Last modified on Nov 06, 2023 16:30

1. Summary

This technical guide summarises a case study for the use of metocean data from the Copernicus Climate Change Service repository, the Climate Data Store (CDS). The study is an assessment of the effect of climate change on European offshore wind operations and maintenance (O&M). Data was used for several locations, for historical (1977-2003), nearfuture (2041-2070) and far-future (2071-2100) epochs. Future data was associated with Representative Concentration Pathways (RCPs) 4.5 and 8.5 climate scenarios, the middle and highest emission cases, respectively (Detlef, et al. 2011).
The metocean data was first analysed to assess the impacts of climate change. It was shown that there was typically a small reduction in wind speeds and wave heights. Comparing the results for the RCP 8.5 far-future scenario, to the historical ones, this related to a 2% mean reduction in mean wind speed and increases in maintenance-related turbine accessibility of 0-5%, suggesting increases in availability.

The O&M assessment was carried out by performing simulations using JBA Consulting's ForeCoast^® Marine software. The development of the O&M modules was informed by engagement with key offshore wind industry parties. Modelling used an offshore wind farm configuration representative of the next generation of wind farms; 100 ten Megawatt (MW) turbines. The Climate Impact Indicators (CIIs) produced by the simulations were monthly and annual Available Energy (energy generated without downtime) and Generated Energy (Available Energy minus missed generation due to maintenance), Turbine Availability and Jack-Up Barge Charter Duration (as an indicator of cost).

The variations in the CIIs were summarised by comparing the 50^th percentile results for the far-future RCP 8.5 scenario, to the results for the historical case. Considering the mean differences across all simulated locations, the changes were reductions of 3% in available and Generated Energy and 2% in jack-up barge usage, with a 0.2% increase in Turbine Availability. The impacts on the future of the offshore wind industry in Europe were assessed using present day strike prices, consumer energy usage and considering the potential scale of European offshore wind generation in the far-future period, 2085. A reduction of 3% in generation equated to a drop of 16 TWh/year, which is equivalent to €1,000 million/year in missed revenue, a reduction of 8 million tonnes CO₂/year in replaced fossil fuel use and 4 million less homes powered.

This study has shown that, in Europe, climate change may impact both the amount of energy produced by an offshore wind farm and the usage of vessels chartered to repair offshore wind turbines. For offshore wind farm operators this may affect the viability of a wind farm or require innovations in terms of turbine design or maintenance strategies. With regards to energy policy in Europe, reductions in available and generated wind power may influence the scale of development as additional turbines or potentially even wind farms may need to be developed to account for the shortfall in renewable energy a decrease in wind speeds could produce. In this regard, strategic decision making can be supported using the results in the CDS for financial calculations or by employing modelling software, like ForeCoast® Marine, to simulate in detail the effects of such changes.

It should be noted that the climate change impact information presented in this report was based on data from a single available climate projection ensemble member. Whilst a more robust study would include analyses run with a range of ensemble members, it does provide, for purposes of this demonstrator, an indication of the possible effects of climate change on offshore wind farm operations.

1.1. Use Case description

1.1.1. Issue to be addressed

This case study examined the potential impacts of climate change on the operations and maintenance (O&M) of offshore wind farms. O&M is the monitoring, management, upkeep and repair required to balance the running costs and electricity generation of a project (GL Garrad Hassan 2013).
Offshore meteorological and oceanographic (metocean) conditions have the potential to affect O&M in the following ways:

• Vessel and maintenance crew access to turbines can be disrupted, delaying maintenance;
• The power produced by wind turbines is a function of wind conditions.

Both of these factors have the potential to impact the energy yield and therefore the revenue produced by offshore wind farms; which in turn may affect the viability of offshore wind energy as a renewable source of electricity.

To understand the extent of these potential impacts, an offshore wind farm O&M logistics model was constructed by further developing and utilising JBA's proprietary software ForeCoast^® Marine. The model has been configured to represent the lifecycle of an offshore wind farm, including power and revenue stream from the turbines as well as modelling turbine failure modes, which require technicians and vessels to carry out repairs.

1.1.2. Decision support to client

The results can be used to show whether mitigation strategies will be required to offset the impacts of climate change and thus allow offshore wind to remain a viable source of renewable energy (and return a profit to developers). To help develop offshore wind farm sites in the present day, the model can also be used to design and optimise O&M strategies.

1.1.3. Temporal and spatial scale

The case study is most relevant for strategic planning for the coming 20 to 30 years, which aligns with the development cycle of a windfarm. It is likely the technology involved in offshore wind farms and the supporting infrastructure, such as vessel and autonomous vehicles, will change significantly beyond this epoch so that modelling of offshore wind O&M beyond this time is likely to be unrepresentative of the industry. Modelling was also carried out for seven wind farm development regions across current and future development and tendering zones in Europe.

1.2. Potential adaptation measures

1.2.1. Lessons learnt

This section summarises the lessons that were learnt when using the Climate Change Service in the workflow used to produce the results in this case study. The step in the workflow that involved use of the service was acquisition of metocean data from the CDS.

The data were used to carry out an assessment of the impact on offshore wind O&M across Europe.

The successful factors related to the metocean data collection were:

• The dataset contained significant wave height, Hs, and wind speed, which are typically the most important variables for this type of modelling (Dinwoodie, et al. 2015).
• Being hourly timeseries, the dataset had sufficiently high temporal resolution to provide the required accuracy for the O&M modelling.
• The locations available in the dataset were sufficiently numerous and widespread to allow a pan-European study, of sites relevant to the offshore wind industry, to be carried out.

The limiting factors related to the metocean data collection were:

• The study presented was carried out for existing or in-development sites. These tend to be relatively close to the coast, for example within 60 km, so the available locations in the wave height dataset, shown in Figure 1, were adequate. In the future, however, there is likely to be significant development further offshore, for example the Dogger Bank site, 130 km off the coast (the Guardian 2015). For these sites, available locations in the dataset will likely not be close enough for accurate O&M modelling.

Figure 1: Map of locations of wave height data.

• Only one timeseries was available for each location and time period/climate scenario. This lack of a population of ensemble members meant that the uncertainty in the metocean data could not be assessed for individual datasets. It was therefore not possible to estimate the uncertainty in the Climate Indicators produced.

1.2.2. Importance and relevance of adaptation

Before this study, the impacts of climate change on offshore wind O&M had not been widely considered. Modelling of offshore wind O&M had previously been carried out using historic data and sensitivity testing undertaken by testing incremental changes in the metocean conditions and measuring the impact on the offshore wind farm (i.e. with respect to Turbine Availability and energy production). The CDS service has allowed O&M models to be run using climate change projections for the first time, which provides an understanding of the severity and an indication of the likelihood of climate impacts.

1.2.3. Pros and cons or cost-benefit analysis of climate adaptation

This study has shown that there could be a 3% reduction in Generated Energy for offshore wind in Europe. This is later equated to a drop of 16 TWh/year, or €1,000 million/year in missed revenue, a reduction of 8 million tonnes CO₂/year in replaced fossil fuel use and 4 million less homes powered. The gains for climate adaptation are hence avoidance of these reductions in these important financial, environmental and energy metrics. The decisionmaking for how this is achieved can be supported by using the results in the CDS for financial calculations or employing modelling software, like ForeCoast® Marine, to assess the effects of changes in variables such as turbine reliability and O&M strategy.

1.2.4. Policy aspects

The results of this study have shown that, within the boundaries imposed by the available metocean data, offshore wind energy can continue to be developed as a means of meeting renewable energy targets and hence reducing carbon dioxide emissions. However, the 3% mean reduction in Generated Energy is useful to consider in policy making, with an equivalent increase in low-carbon energy production required to match the shortfall.

1.3. Data production and results

The stages that were completed to generate the results for this case study are described in this section and summarised in Figure 2.

Figure 2: Workflow of case study results generation.

1.3.1. Step 1: Metocean data collection

Significant wave height, Hs, and wind speed data from the CDS were used as inputs to this study. The significant wave height data was from the Tier 1 data, while the wind speed data was from the HIRHAM5/EC-EARTH model that was used as an input for the Tier 1 data creation. The data was extracted for seven key locations around Europe, shown in Table 1.

A Python script was created that calculated the distance between the desired location and all the data-points in the wind and wave data-sets. For each data-set, the data-point located at the shortest distance from the desired location was selected. Data for this location – and the specified time period – were then extracted from the CDS NetCDF data file. Wind data were provided in terms of the eastward (uas) and northward (vas) nearsurface wind. Thus, it was necessary to calculate the wind speed magnitude by applying Pythagoras' theorem.
The simulations were run for the locations shown in Table 1 and Figure 3, which were representative of existing or planned offshore wind farms.

Table 1: Locations used for O&M modelling

Figure 3: Map showing locations used for O&M modelling.

For each location, the climate scenarios shown in Table 2 were used.

Table 2: Climate scenarios used for O&M modelling

1.3.2. Step 2: Wind speed conversion

The wind speed data was modelled at 10 m above sea level. In the O&M model, the height of the wind turbine hubs and height of the jack-up barge crane, used for lifting operations, were assumed to be at 120 m and 150 m, respectively. Within the boundary layer above the air-water interface, wind speeds increase with height. Hence, the wind speeds were converted to these heights and used for finding weather windows for the lifting operations and calculation of power production, respectively. This was achieved by applying the standard power-law formula, shown in ( 1 ).

Where:
𝑈(𝑧) is the wind speed at height, z;
𝑈(𝐻 )is the wind speed at reference height, H;
𝛼 isthepower-lawexponent,0.1,a standard value for the open sea (DNV2007).

1.3.3. Step 3: Risk factor analysis

Before the metocean data was used in the modelling, it was analysed to help predict the effects of climate change on O&M. The two primary variables in offshore wind that are affected by metocean conditions are the wind power production and the availability of the wind turbines.

Wind turbines produce power by converting the kinetic energy in the wind to first mechanical energy, in the rotor and drive-train, and then to electrical energy, in the generator. The input wind power available for this conversion, P, is calculated using ( 2 ).

Where:
𝜌 is the air density;
𝑈 is the wind speed;
𝑑 is the diameter of the turbine.

The key point to note from ( 2 ) is that the available power is proportional to the cube of the wind speed. This means that a small change in the wind speed at a wind farm site could in fact have a large impact on the available power. A 10% increase in wind speed, for example, would result in a 33% increase in available power.
The actual power generated by a turbine is related to the wind speed via a 'power curve', with the one used in this study shown in Figure 4.

Figure 4: Power curve used in modelling (Dinwoodie, et al. 2015).

Figure 4 shows that the relationship between generated power and wind speed isn't as simple as that shown in ( 2 ), due to various factors related to efficiency and structural integrity. This means that a change in wind speed won't necessarily lead to that change cubed. However, if the commonly occurring wind speeds at a site fall into the period where the turbine starts to produce power, between the 'cut-in' speed, at 3 m/s, and the point when the power plateaus, at its 'rated output', at 17 m/s, then the power will be dependent on the speed. This is because the relationship between generated power and wind speed is cubic for this period and therefore highly sensitive. Hence, a change in the distribution of wind speeds in this range will result in a significant change in generated power and hence revenue.

The availability of a wind turbine is the actual time that it operates (for 'time-based' availability) or the energy that it produces (for 'energy-based' availability) divided by the maximum that it could have achieved without experiencing downtime. Downtime occurs due to corrective or preventative maintenance requiring the turbines to be turned off and hence stop producing power. This maintenance is typically performed by technicians, transferred to the turbines by transport vessels, and heavy lift vessels (GL Garrad Hassan 2013). All of these vessels usually require metocean conditions to be below certain thresholds to access the turbines. Hence, changes to the metocean conditions will impact the turbine accessibility, thus altering the timeliness of when turbine maintenance is performed and finally a change to the turbine availabilities. The specific variables that will likely impact turbine availabilities in this study are defined in Table 2.

Table 2: Metocean thresholds for operations for assessing vessel turbine accessibility.

From the above risk analysis, the metocean variables of primary concern for the industry are the wind speed and significant wave height.
The risk due to changes in wind speed and significant wave heights was first evaluated by plotting cumulative distribution functions of the data for each scenario and overlaying them for each location. Example graphs for a typical location are presented in Figure 5 and Figure 6 for wind speed and significant wave height, respectively. Also shown are the metocean thresholds, defined in Table 2.

Figure 5: Cumulative distribution function of 150 m height wind speed and metocean thresholds for affected operations, for each climate scenario for a typical location ('Northern North Sea').

Figure 5 shows that, between the turbine cut-in and rated output speeds, there is an, albeit subtle, trend of reducing values for the wind speeds going from the historical data to the future and higher emission scenario data. This is shown by the future curves falling left of the historical curves. This is significant because approximately 80% of the values fall within this range and hence a reduction in wind speeds in this range will likely result in a noticeable decrease in mean available power.
For the 'lifting in major repairs' operation, the reduction in wind speeds can be seen by the fact that the values on the y-axis for the 10 m/s value on the x-axis increase, indicating an increasing chance of values being below this threshold. This indicates that the jack-up barge will have higher 'accessibility' for performing this operation.

Figure 6: Cumulative distribution function of significant wave height, Hs, and metocean thresholds for affected operations, for each climate scenario for a typical location ('Northern North Sea').

Figure 6 shows that, for all operations, due to reductions in significant wave height, there is a trend of increasing turbine accessibility going from the historical results to the future and higher emission scenario data.

The results so far have been for a single representative location. The risk is considered further and generalised by assessing changes in the metocean conditions across the locations. This is achieved by using indicators for the wind speed and significant wave height related to the available power and turbine accessibility, respectively. The wind speed indicator is the mean value; the indicators for turbine accessibility are the percentages of time that the metocean conditions were below the thresholds required for operations required for turbine maintenance, as defined in Table 2. The changes in variables were assessed by computing the percentage differences between the values for the most extreme case, the far-future RCP 8.5 scenario, and the historical results. This was carried out for each location and then the minimum, mean and maximum values were calculated, which are presented in Figure 7.

Figure 7: Changes in metocean risk indicators between far-future RCP 8.5 results and historical results.

Figure 7 shows that there is expected to be a reduction in mean wind speeds for all locations. The mean reduction was 2%, which, using ( 2 ), equates to a 6% reduction in available power. This suggests that offshore wind farms are at risk of generating less energy and hence revenue. In contrast, the consistent mean increases in accessibility indicate that the vessels will be able to perform the operations required for maintenance with less time waiting for weather windows. This suggests that maintenance will be performed more quickly, getting turbines online faster and hence increasing availability. This could counteract the drop in available power.

1.3.4. Step 4: Model set-up

The risk analysis performed in section 4.4 gave an indication of the impacts of climate change on offshore wind. However, they were quite simplistic, not considering the interplay between factors such as vessel availability, wind turbine failure rates and the timing of weather windows. Time-domain modelling of offshore operations allows these relationships to be considered (Beamsley, et al. 2007). The Gamer Mode module of ForeCoast^® Marine (FCM), an award winning advanced marine logistics modelling software is such a system (JBA Consulting 2019). ForeCoast^® Marine models marine operations by simulating the workflow of the vessels, technicians and other resources. The software evaluates delays by identifying weather and operational windows using a time series of metocean variables that impact the operations, in this case the wind speed and wave height.

This is repeated many times over the extent of the timeseries, generating a rich ensemble of results, used for statistical analysis.
The model represents all the turbines, and defined component parts. If the wind is blowing at a given time, the turbines will generate power and revenue according to the wind speed and a defined power curve and revenue model. It is important to note that is not a precise representation of revenue generation (as is does not account for turbine wake effects, for instance), but it is of value in relative terms.

As the simulations progress, the model 'orders' scheduled maintenance operations according to a define programme or frequency. It also injects random failures according to the failure rates/modes inputted (which is a function of the components). When a maintenance operation is required, the model's scheduler will order the repair method, including the types of vessels, technicians and plant needed. If these resources are available (i.e. not be used on another operation), a maintenance activity will be attempted. However, if the weather does not allow this, the vessels will be held at port and lost revenue and vessel costs will accumulate. The image below illustrates a screen shot demonstrating these key features.

To model the O&M of offshore wind, additional Gamer Mode modules were developed to simulate turbine failures, schedule repair tasks, charter additional vessels, plan preventative maintenance tasks and generate energy and revenue. Data analysis modules were also written to calculate the key performance indicators of the wind farm, including energy yield, revenue, lost revenue, O&M costs and Turbine Availability.

1.3.5. Step 5: Model verification

The first stage in the simulation process was to confirm that the model functioned correctly. Without access to long-term operational offshore wind data for validation, the next best approach for this was applied. This was to 'verify' the model through comparison of results produced by it to those from four other models presented in a paper (Dinwoodie, et al. 2015).
The verification model consisted of 80 three MW turbines, operating in the southern North Sea. The distance from the O&M port to the farm was 50 km and the distance between turbines was 1 km. The input weather data was from the FINO1 research platform (FuEZentrum FH Kiel GmbH 2018).
A 'base case' O&M strategy was defined, with further variations on this simulated to confirm that the model behaved as expected. The base case consisted of:

• 6 types of maintenance – 5 corrective types (manual resets, minor repairs, medium repairs, major repairs and major replacements) and 1 preventative type (annual services);
• 3 crew transfer vessels (CTVs), 1 field support vessel and 1 heavy lift-vessel (HLV);
• 20 technicians.

The results from the FCM model were compared to the average, minimum and maximum results from the paper. The variables that were used for comparison were average TimeBased Availability and annual total O&M cost, with results presented in Figure 8 and Figure 9, respectively.

The results in Figure 8 and Figure 9 show that the FCM model produced results that were generally very close or within the ranges of results from the other models and were typically above the mean value. These consistently optimistic results suggest that there were slight differences in either the model inputs or in the modelling techniques, or a combination of these factors. For example, the scheduling of tasks, vessels and technicians in the FCM model is achieved using vehicle routing algorithms, which may account for the improved efficiency and hence increased Turbine

Availability seen in Figure 8; although it is not possible to draw firm conclusions without access to the models described in the literature (Dinwoodie, et al. 2015).

The key outcome of the verification was that the FCM model responded in a similar manner to the models represented in the paper when varying the model parameters. Applying increased failure rates, for example, resulted in a comparative reduction in availability and an increase in costs. This verified response of the model to changes in input parameters was essential to this study, which has investigated how changes in metocean conditions could impact the overall performance of an offshore wind farm. Following this verification study, it was considered that the FCM O&M model was suitably tested and shown to be a valid model for this Use Case.

Figure 8: Time-Based Availability results for O&M model verification.

Figure 9: Direct O&M cost results for O&M model verification.

1.3.6. Step 6: O&M simulations

Simulations were carried out for a hypothetical offshore wind farm, typical of the North Sea and with a bias towards those that will be operational from the 2020s onwards (4C Offshore 2019). The farm consisted of 100 ten MW turbines. The maintenance requirements, vessel metocean limits and wind turbine power curves were drawn from the same literature as used in the model verification, section 4.5 (Dinwoodie, et al. 2015).

The O&M strategy of an offshore wind farm encompasses the choices of numbers and types of resources (technicians and vessels) used to maintain the turbines, the hours in which they work and the port(s) that they travel to the farm from. The O&M strategy for the wind farm was chosen to partly reflect the vessel technology likely utilised for the next decade at least. Even if the chosen technology becomes outdated within the time period for the climate projections, the relative changes in O&M performance are still relevant. The other requirement for the strategy was that it achieved a typical availability for the wind turbines (please see section 4.7 for an explanation of availability calculation) of approximately 95% in the current day climate (SPARTA 2018).

The vessel strategy chosen was to use one service operation vessel (SOV) and two crew transfer vessels (CTVs) to transport technicians to carry out minor-medium repairs and one jack-up barge to perform major repairs. In the simulations, the SOV stayed offshore for 2 weeks, transferred technicians in significant wave heights, Hs, of up to 3 m and had a capacity of 60 technicians, while the CTVs stayed offshore for 12 hours, transferred in Hs of up to 1.5 m, and each had a capacity of 12 technicians. The jack-up barge had no limit on the time that it stayed offshore, but was restricted to jacking up and down when Hs was below 2 m and only performing lifting operations when the wind speed was below 10 m/s. All transport vessels carried their maximum numbers of technicians. The transport vessels and technicians were permanently employed by the offshore wind farm, while the jack-up barge was chartered for two months each time that it was required.

The model was run for each year in the datasets, with the length of each O&M campaign being five years. The outcome of this was that each simulation produced results for around 30 O&M campaigns.

1.3.7. Step 7: Data analysis

During the simulations, the statuses (locations, times, failure modes, etc) of the vessels, technicians and wind turbines were recorded. Combined with the wind speed data and the power curve of the wind turbines, this information was processed to produce the O&M Climate Indicators. For each Indicator, the monthly and annual/campaign values were calculated. The only exception to this was the jack-up barge usage, which was only calculated annually. From the O&M campaign iterations, for each variable, the P10, P50 and P90 percentiles were calculated.

Table 3 provides a summary of the offshore wind climate indicators, including their names, descriptions and equations used for their calculation. The presented variables will also be available to the users of the CDS for combining with their inputs, such as electricity unit strike price, for calculation of other results, for example revenue.

Table 3: Climate Indicators – names, descriptions and equations.

1.3.8. Step 8: Results

A sample of the results available to the user in the CDS are presented here. The results are both for a typical location, with the chosen representative location being the 'Northern North Sea', as well as summary comparative results for all locations. Note that the graphs shown are not in the exact same format as those that will be displayed in the CDS, but they are similar. Additionally, not all results provide error bars and hence do not reflect some of the uncertainty in the values. Shown first are monthly results, followed by annual values.

1.3.8.1. MonthlyAnalysis

Shown in Figure 10 is an example plot of monthly Available Energy for the example location.

Figure 10: Monthly mean available and Generated Energy for historical time period for example location ('Northern North Sea').

Figure 10 shows the expected behaviour of greater levels of energy being available and generated in the winter months, due to higher wind speeds.

Figure 11 is an example plot of monthly Energy-Based Availability for the example location. The expected behaviour of increased availability in the summer, due to lower wind speeds and wave heights allowing greater accessibility to the wind turbines. The significant drop in availability in May was due to annual preventative maintenance, which requires the wind turbines to be shut down, being carried out then.

Figure 11: Monthly mean Energy-Based Availability for historical time period for example location ('Northern North Sea').

1.3.8.2. Climate Change Scenario Comparison

Shown below, are the plots for the annual values for different time periods and climate projections for the example location. Figure 12 shows the results for Available Energy – the amount of energy produced if the wind farm was operating at 100% availability. The P50 annual Available Energy is likely to reduce in the future, with this study showing a maximum decrease of 4% relative to the historical value, occurring for the far-future RCP 8.5 scenario. This is due to the small decrease in wind speeds, as shown in Figure 5.

Given that there is a decrease in Available Energy due to lower wind speeds, it is reasonable to predict that there will be an increase in Turbine Availability as the significant wave height drops with wind speed and the vessels are able to access and repair turbines more readily. This hypothesis is also supported by considering the increases in turbine accessibility in Figure 5 and Figure 6. Indeed, Figure 13 shows that Energy-Based Availability does increase, however the change is marginal. For example, there was an increase of 0.2% from the P50 historical and RCP 8.5 far-future results. The same result was found for the Time-Based Availability.

Figure 12: Comparison of annual Available Energy for historical, near-future and far-future for RCP climate scenarios for example location ('Northern North Sea').

Due to the negligible change in Turbine Availability and the fall in wind speed, the Generated Energy results displayed the same trend as the Available Energy, culminating in a 4% reduction between the far-future RCP 8.5 results and the historical values Figure 14.

Figure 13: Comparison of mean Energy-Based Availability for historical, near-future and far-future for RCP climate scenarios for example location ('Northern North Sea').

A reduction in Generated Energy would have a direct impact on the revenue at this location. However, losses in revenue may be offset by a reduction in the cost of vessels. The expenditure for jack-up barges typically represents the largest cost for O&M and is proportional to the number of days that the vessels are chartered for (Dinwoodie, et al. 2015).

Figure 15 shows that there is likely to be no appreciable change in the annual chartering durations for the jack-up barges used for major repairs. Due to reductions in wind speed and significant wave height, shown in Figure 5 and Figure 6, respectively, it is likely that vessel operability will increase for the future scenarios and allow improved access to offshore wind turbines. However, it has been shown here to be unlikely to influence the charter duration. This is partly because the types of vessels are usually chartered for fixed durations that have contingency time built in. Use of the metocean data in the CDS for a more detailed study would allow users to potentially reduce the durations that they charter jack-up barges for and hence save money.

Figure 14: Comparison of annual Generated Energy for historical, near-future and far-future for RCP climate scenarios for example location ('Northern North Sea').

Figure 15: Comparison of annual jack-up barge usage for historical, near-future and far-future for RCP climate scenarios for example location ('Northern North Sea').

Similar results were found for the other locations considered. The variations in the Climate Impact Indicators were summarised by comparing the 50^th percentile results for the RCP8.5 far-future scenario to the results for the historical case. The mean, minimum and maximum changes were calculated and are presented in Figure 16.

Considering the mean results shown in Figure 16, the changes were reductions of 3% in available and Generated Energy and 2% in jack-up barge usage, with a 0.2% increase in Turbine Availability. All results are presented in Table 5 in the appendix.

Figure 16: Changes in P50 Climate Impact Indicators for all locations between RCP 8.5 far-future results and historical results.

1.3.8.3. Assessing the Impact on the Offshore Wind Industry in Europe

The 3% reduction in Available Energy, which did not vary significantly around Northern Europe, occurred due to a general drop in wind speeds, as indicated by the 2% reduction in the mean wind speed, shown in Figure 7. Due to the increases in vessel operability, also shown in Figure 7, resulting from reductions in wind speed and significant wave height, it was expected that there would be an increase in Turbine Availability. However, the results produced here have shown that the relationship between turbine accessibility and Turbine Availability is not necessarily linear. The increase in Turbine Availability was small and not large enough to offset the reduction in Available Energy. This resulted in the decrease in Generated Energy being the same as for the Available Energy.
Due to the size and potential of the Offshore Wind industry in Europe (Wind Europe and BVG associates 2017), a 3% change in Generated Energy could have significant implications. The impact of this is estimated by predicting the absolute change in annual European offshore wind energy generation and equating it to different units.
The first step is to estimate the annual European offshore wind energy generation in the middle of the far-future period, i.e. 2085, 𝐸_'(. This is calculated by applying Equation 3.

Where:
C is the current average capacity factor;

I is the predicted total installed capacity by 2085;

y is the number of hours in a year, 8760

An estimate for C was 38%, which is the lifetime mean of the capacity factors, as of 2019, in the UK – the country with the highest amount of installed capacity (Energy Numbers 2019); I was estimated by taking the current offshore installation rate, 2.6 GW/year (the average of the last four years, which have stabilised), multiplying it by the years between now and 2085, 66, and adding this to the current installed capacity, 18.5 GW, to give a total of 190 GW (Wind Europe 2018). Note that this is a reasonable value, given that the economic potential of Europe has been estimated as 607 GW (Wind Europe and BVG associates 2017). This means that the predicted annual energy generation in 2085 is 630,000 GWh/year, i.e. 630 TWh/year. A 3% reduction in this equates to 16 TWh/year. The significance of this can be enhanced through conversion into other units, as has been carried out in Table 4.

Table 4: Equivalent values for reduction in Generated Energy for far-future RCP 8.5 scenario compared to historical results.

The results in Figure 16 show that a reduction of 3% in Generated Energy by offshore wind could have significant consequences for the European economy, meeting environmental targets and supplying electricity generated from renewable sources.

Due to its dependency on metocean conditions for both jacking up/down and lifting components for repairs, usage of the jack-up barges was shown to be more sensitive to climate change than Turbine Availability, as indicated by the 2% mean reduction. The relatively wide range of results is likely most related to insufficient model convergence, rather than geographical dependency. However, the mean reduction is good news for wind farm operators, due to an expected decrease in the costs associated with chartering jackup barges.

One way that the loss of revenue, shown in Table 4, could be offset is a reduction in the expenditure required for chartering jack-up barges, resulting from the 2% mean reduction in length shown in Figure 16. Using the mean charter length for the historical results, 206 days/year, a 2% reduction equates to 3.5 days. Using a charter rate of €180,000/day, this is equivalent to a saving of €0.6 million/year per 1 GW wind farm (Dinwoodie, et al. 2015). Scaling this to the projected installed capacity of 190 GW, this equates to a saving of € 120 million/year.

Using the climate change projections produced for C3S, these results have shown that, due to climate change, there may be small reductions in wind farm energy production and jack-up barge usage and no appreciable change in availability. While unlikely to require dramatic changes to turbine technology or O&M strategy, the reduction of energy generation should be compensated for, for example with improvement in turbine efficiency or reliability or increasing the European installed capacity. The decision-making for how this is achieved can be supported by using the results in the CDS for financial calculations or employing modelling software, like ForeCoast® Marine, to simulate the effects of such changes.

1.4. Conclusion

Wind speed and significant wave height hourly timeseries data were downloaded from the CDS for 'historical' (1977-2003), 'near-future' (2041-2070) and 'far-future' (2071-2100) time periods. The future periods were for RCP scenarios 4.5 and 8.5, the middle and highest emission scenarios, respectively (Detlef, et al. 2011). This data was used as a key input for modelling of offshore O&M in sites around Europe, relevant to constructed farms and near-future development.

The metocean data was analysed to predict the impacts of climate change by comparing the distributions of data for all time periods and climate scenarios. It was shown that there was typically a small reduction in both wind speeds and wave heights. This resulted in a 2% mean reduction in mean wind speed and increases in maintenance-related turbine accessibility of between 0% and 5%, suggesting increases in Turbine Availability.
Guided by engagement with partners in the offshore wind industry, a model was developed to simulate the primary factors involved in O&M and to produce information required to generate Climate Impact Indicators. Before using the developed model for this case study, it was verified through comparison of results to those available in the literature (Dinwoodie, et al. 2015).
The O&M model was used to simulate a typical wind farm running for each of the locations and time periods/climate scenarios. The variations in the Climate Impact Indicators were summarised by comparing the 50^th percentile results for the highest emission and furthest ahead period, the RCP 8.5 far-future scenario, to the results for the historical case. Considering the mean differences across all the locations, the changes were reductions of 3% in available and Generated Energy generation and 2% in jack-up barge usage, with a 0.2% increase in Turbine Availability.

Using today's conversion figures and predicting the scale of European offshore wind generation in the middle of the far-future period, 2085, the key predicted changes in Climate Impact Indicators were equated to absolute values. A 3% reduction in generation was equated to a drop of 16 TWh/year. This is equivalent to €1,000 million/year in missed revenue, a reduction of 8 million tonnes CO₂/year in replaced fossil fuel use and 4 million less homes powered. The 2% reduction in jack-up barge charter use was equated to a saving of € 120 million/year, or only 12% of the lost generation revenue.

This study has shown that, in Europe, climate change may impact both the amount of energy produced by an offshore wind farm and the usage of vessels chartered to repair offshore wind turbines. For offshore wind farm operators this may affect the viability of a wind farm or require innovations in terms of turbine design or maintenance strategies. With regards to energy policy in Europe, reductions in available and generated wind power may influence the scale of development as additional turbines or potentially even wind farms may need to be developed to account for the shortfall in renewable energy a decrease in wind speeds could produce. In this regard, strategic decision making can be supported using the results in the CDS for financial calculations or by employing modelling software, like ForeCoast® Marine, to simulate in detail the effects of such changes.

It should be noted that the climate change impact information presented in this report was based on data from a single available climate projection ensemble member. Whilst a more robust study would include analyses run with a range of ensemble members, it does provide, for purposes of this demonstrator, an indication of the possible effects of climate change on offshore wind farm operations.

2. Appendix

Table 5: Percentage differences between P50 Climate Impact Indicators for RCP 8.5 far-future scenario results and historical results for all locations.

3. References

4C Offshore. 2019. Hornsea Project One Offshore Wind Farm. https://www.4coffshore.com/windfarms/hornsea-project-one-united-kingdom-uk81.html.

Beamsley, Brett, Peter Mccomb, David Johnson, and J. Silk. 2007. "Estimating weather downtime for ocean engineering using Sequential Downtime Analysis (SDA)." 18th Australasian Coastal and Ocean Engineering Conference 2007.

Detlef, P. van Vuuren, Jae Edmonds, Mikiko Kainuma, Keywan Riahi, Allison Thomson, Kathy Hibbard, George C. Hurtt, and etc. 2011. "The representative concentration pathways: an overview." Climatic Change.

Dinwoodie, Iain, Ole-Erik V. Endrerud, Matthias Hofmann, Rebecca Martin, and Iver Bakken Sperstad. 2015. "Reference Cases for Verification of Operation and Maintenance Simulation Models for Offshore Wind Farms." Wind Engineering.
DNV. 2007. "Environmental conditions and environmental loads DNV-RP-C205." Energy Numbers. 2019. UK offshore wind capacity factors. Accessed March 14, 2019.

FuE-Zentrum FH Kiel GmbH. 2018. FINO1 research platform. https://www.fino1.de/en/.

GL Garrad Hassan. 2013. "A Guide to UK Offshore Wind Operations and Maintenance." http://www.hienergy.org.uk/Downloads/General%20Documents/guide-to-uk-offshore-wind-operationsand-maintenance.pdf.

JBA Consulting. 2019. ForeCoast(R) Marine. https://www.forecoastmarine.com/.

renewable UK. 2018. Wind Energy Statistics Explained. Accessed March 2019, 14. https://www.renewableuk.com/page/UKWEDExplained.

SPARTA. 2018. "System Performance, Availability and Reliability Trend Analysis – SPARTA 2017/18 Portfolio Review."

Teyabeen, Alhassan, Fathi Akkari, and Ali Jwaid. 2017. "Power Curve Modelling For Wind Turbines." UKSim-AMSS 19th International Conference on Modelling & Simulation.

the Guardian. 2015. World's biggest offshore windfarm approved for Yorkshire coast. Accessed February 18, 2019. https://www.theguardian.com/environment/2015/feb/17/worldsbiggest-offshore-windfarm-approved-for-yorkshire-coast.
Wind Europe and BVG associates. 2017. "Unleashing Europe's offshore wind potential A new resource assessment." https://windeurope.org/wp-content/uploads/files/aboutwind/reports/Unleashing-Europes-offshore-wind-potential.pdf.

Wind Europe. 2018. "Offshore Wind in Europe Key trends and statistics 2018." https://windeurope.org/wp-content/uploads/files/about-wind/statistics/WindEuropeAnnual-Offshore-Statistics-2018.pdf.

Wind Power Offshore. 2018. UK offshore capped at €63/MWh for next auction. Accessed March 2019, 14. https://www.windpoweroffshore.com/article/1519177/uk-offshore-capped%E2%82%AC63-mwh-next-auction.

4. Related articles

• Page:
  NEMO-ERSEM and POLCOMS-ERSEM: User Guides for Products (v1.21)
• Page:
  NEMO-ERSEM and POLCOMS-ERSEM: User Guides for Products (v1.1)
• Page:
  Summary indicators (Tier 1): Product User Guide
• Page:
  Risk indicators (Tier 3): Product User Guide
• Page:
  Loss indicators (Tier 3): Product User Guide