We’re looking at ways in which we can install the foundations for a North Sea offshore wind farm and using Mermaid to make decisions on the vessels and the strategy which best suit our needs. In the previous post in this series we discussed the base case for this analysis process, this being our first look at how we might perform the work.
In this post we’re going to take a look at the results of the analysis we ran last time and we’ll try to work out how, and indeed if, we can improve things.
The Simulations
We ran the installation of 10 foundations and of 72 foundations using start dates across the entire year, this lets us develop knowledge around the impact of starting at different times of year. Let’s take a look at the results.
The Results
We can generate a number of different outputs in Mermaid. Shown below is the cost duration box-whisker plot for the installation of 10 foundations. This chart groups runs together by start date generating data sets for beginning the operations in each month of the year. The boxes represent the interquartile range, the horizontal line inside each box the median and the whiskers the maximum and minimum durations. Where durations are unknown, for example when the end of the metocean data is reached and an analysis can’t complete, the maximum whisker shoots off the top of the chart; this is seen in November and December. The solid horizontal line represents the unweathered case, this being the perfect installation which cannot be bettered.
In this chart we can see the typical seasonality we might expect to get, with operations starting in April, May and June being the quickest and least risky (we can assess risk from the spread of the whiskers and the size of the interquartile region, smaller being lower risk). If we look at the data for March, the median installation duration is about 1000 hours, or 42 days, meaning that work starting in March runs through to mid- to late-April. Operations starting in November are particularly impacted by weather, with the operations taking place through the winter (a quick note, we’re using northern hemisphere winter and summer). Operations starting in October and January are fairly risky as the maximum and minimum lie a long way outside the quartile range. The longest duration in October is about 2500 hours, or 125 days, meaning that operations starting in October may not complete until the start of the following spring.
Now let’s take a look at the same chart for all 72 foundations, it’s shown below. Hang on! What’s going on here?
Firstly, there are a lot of unknown maximum values. This is understandable as this is a long installation process and a number of the runs towards the end of the 10 year metocean set hit the end of the data – let’s ignore the maximum values as we analyse this chart.
The winter start dates have shorter durations than the summer start dates. This is the exact opposite of the trend we saw in the previous chart and not what we would expect; why is this occuring? Remember that 1 year is 8760 hours. So a most of our median values lie just under 1 year at about 11 months. This means that an installation starting in December will be finished by the end of October and we won’t try to work in the poor weather seen in November as we’re done. Basically, the 11 month weathered installation time means that we knock one of the months out of the working time and that this is the month before the one we start in. Looking at this a little more: If we start our operation in May the month we don’t work in is April, which we know is a good month from our previous chart and our progress is affected.
The seasonality trend, whilst inverse, is much flatter than previously. This is once again due to the weather conditions being experienced during work. Now most of our starts incur a full year of weather, both good and bad, and therefore the variation between the starts is reduced.
The Next Steps
So we’re done, right? Start the operation in January and there’s a pretty good chance we’ll have all our foundations in by the end of the year? We could do this but we know from our first simulation that the spring and summer months are much better for us. Could we better exploit this good weather rather than incurring winter downtime that we could avoid (remember the winter start is better because we avoid working in one of the winter months, not because the winter months are easier to work in)?
To improve (shorten) our installation duration, reduce our cost (as a result, possibly) and get our wind farm installed and generating revenue sooner (assuming the turbines go on the foundations successfully) we could do a few different things:
- Get more equipment on board the vessel: We can only carry one foundation and one transition piece on our current vessel. If we can carry more, or find a vessel with larger carrying capacity, we can reduce the transit we do. This means the time at sea can be spent installing foundations rather than travelling back and forth to port.
- Split our installation in half: If we install half one summer and half the next summer we can incur less winter downtime. We’ll need to off-hire the vessel in the winter to keep the price down.
- Bring the foundations to the installation vessel at site: as with carrying more components, we can increase the working time by transiting less. If another vessel (a barge and tugs) brings components to the main vessel, the transit requirements are reduced.
- Use two vessels: Using two installation vessels may speed up the installation by dividing the work. This might let us get everything in in one summer season.
We’ll take a look at using Mermaid to analyse these options in some forthcoming posts, and we’ll consider the thought process and decision making which goes on in support of this analysis and as a result of it.