Back to the programme printer.gif Print




Delegates are invited to meet and discuss with the poster presenters in this topic directly after the session 'Advanced operation & maintenance' taking place on Thursday, 13 March 2014 at 11:15-12:45. The meet-the-authors will take place in the poster area.

Rudy Magne MeteoGroup, France
Co-authors:
Rudy Magne (1) F P Gorka Pérez-Landa (2) Pilar Orellana (2)
(1) MeteoGroup, Best, France (2) IBERDROLA Engineering and Construction, Madrid, Spain

Printer friendly version: printer.gif Print

Presenter's biography

Biographies are supplied directly by presenters at EWEA 2014 and are published here unedited

Rudy Magne holds a PhD in physical oceanography from the Toulon University and a marine science master degree from the ISITV. He worked for 3 years as a researcher for the French Naval hydrographic center in wave modelling and in project management. He joined MeteoGroup in 2011 to conduct oceanographic research and development in MeteoGroup's R&D department.

Abstract

The impact of mesoscale modelling on offshore operational planning

Introduction

Wave height is one of the key metocean elements that dictates whether operational work can be carried out on offshore wind farms. It is also an element that is difficult to correctly forecast in the nearshore environment where wind farms are often located. With the objective of increasing cost-efficiency in marine operations, a study has been conducted into improving the accuracy of wave height forecasts. The study resulted in reducing the forecast error for wave height by up to 45% and subsequently increased the number of potential working hours for wind farm operations and maintenance by over 20%.

Approach

In order to reduce the forecast error for wave height, the impact of mesoscale modelling on wave height forecasting was investigated. Thus, the project team developed a wave forecasting system for the study based on the WaveWatchIII (WW3) model [2] and WRF mesoscale model [2]. In order to quantify the benefits of this approach, a comparison was then made between using GFS global model run instead of WRF to drive the wave forecasting system.

The system included three domains: Atlantic 30', European 10' and Irish Sea 3'x2' nested in a two way approach. Each domain was forced by WRF surface winds (with GFS 0.5° inputs). The WRF simulation featured two-way interactive nested domains at decreasing horizontal grid spacing of 36km, 12km and 4km respectively. The coarser domain extended over the North Atlantic Sea and Western Europe and the inner domain over the UK and Ireland.

To test the added value of WRF in the wave forecast system, the study was carried out with GFS-driven forecasts for each WW3 domain. The system was tested over a two month period between January and February 2013 in “analysis” mode: WRF and GFS analyses and short term forecast (<6h) were used to drive the wave models. Comparisons were performed for several in-situ stations including Liverpool Bay, Cardigan Bay and the Bristol Channel.

The study was set up by IBERDROLA Engineering & Construction (IEC) as an internal R&D project called "Enviroffshore". The goal was to support environmental monitoring and control activities, as well as supporting marine operations according to the needs of future offshore wind farm projects. The IEC's previous knowledge of environmental aspects and atmospheric forecasting capabilities was essential for this development. IEC used MeteoGroup's experience of meteocean forecasting and in the fine-tuning of the WW3 wave model which is coupled to the WRF atmospheric model in the context of the "Enviroffshore System". A pilot case was selected in the Irish Sea, where IBERDROLA, in a joint venture with DONG, are building and will operate the offshore wind farm of West of Duddon Sands (see Figure 1).


Main body of abstract

The behaviour of the WRF model simulating the wind field over the ocean was evaluated against observations from a meteorological masts located 15 km offshore in the surroundings of the wind farm during the selected period. The WRF model simulation was very good with a correlation to the observations of 0.83 (see Figure 2 as an example). This positive result enabled the project team to confidently use the simulated wind fields as the atmospheric forcing for the WW3 model.


Figure 1: Location of the West of Duddon Sands (WoDS) offshore wind farm in the Irish Sea


Figure 2: Scatter plot of observed (X) and simulated by WRF (Y) wind speed (m/s) in meteorological mast location


To validate the WW3 model, the output was first compared to offshore stations (west of France and Ireland) to make sure that the waves entering the Irish Sea were correctly modelled. The output was found to be very accurate with errors (NRMSE) of about 10% for the significant wave height for both GFS and WRF-driven WW3. Over the Irish Sea the errors were found to be larger and more sensitive to both models but with significantly smaller errors when WW3 was driven by WRF; 55% with GFS compared to 30% with WRF.

Figure 3 illustrates the dramatic underestimation of the wave height at Liverpool Bay when the model was driven by GFS. On 30th January the observed wave height at 4pm was 3.4m but the GFS-driven WW3 forecasted just 1.7m wave height. Figure 4 illustrates the comparative accuracy of the WRF-driven WW3 which forecasted a wave height of 3.5m.


Figure 3: Significant wave height driven by GFS winds


Figure 4: Significant wave height driven by WRF winds

Wave height in the Irish Sea and, in particular in Liverpool Bay, is mainly influenced by local wind-wave generation even when long swells occasionally enter the bay. The accuracy of the surface wind speed forecast is therefore crucial in order to correctly resolve the wave growth. The results, as illustrated in Figure 4, demonstrate that WRF correctly rescales the waves even when the wave heights peak.

In addition to the development and testing of the system itself, the study also investigated the potential benefits of the findings. Based on typical operating limits [3], the study calculated the total potential working hours in a one month period (14th January – 14th February 2013) using the WRF-driven system versus the GFS-driven system.

Several maximum wave heights were selected (1.5m, 2m and 2.5m) to reflect the thresholds at which different operational tasks can be carried out eg jacking, anchoring and towing. Over the one month period, the potential working hours based on wave height observations were as follows:
• 512 hours for waves below 1.5m
• 594 hours for waves below 2m
• 651 hours for 2.5m.

The wave system driven by GFS forecasted:
• 620 hours for waves below 1.5m
• 669 hours for waves below 2m
• 688 hours for waves below 2.5m.

The wave system driven by WRF forecasted:
• 464 hours for waves below 1.5m
• 580 hours for waves below 2m
• 634 hours for waves below 2.5m.

The GFS-driven forecasts dramatically overestimated the potential working hours for all three thresholds. In reality this would result in large periods of downtime when scheduled operational work would not be able to go ahead. The WRF-driven forecasts on the other hand are significantly closer to the observations; the gain in working hours by using WRF instead of GFS in the one month period were 60 hours, 61 hours and 20 hours for the respective thresholds.

The same analysis was performed for the Bristol Channel and Cardigan Bay. The results, summarised in Figure 5, shows that between 20-75 potential working hours were gained when using the WRF-driven model compared to the GFS-driven model. The gain decreases with the wave height thresholds, directly related to the local wave climate.


Figure 5: Workable hours gains between GFS and WRF driven WW3


Conclusion

The use of a mesoscale atmospheric model (WRF) has been shown to be crucial to correctly represent the wave field in the Irish Sea and, in particular, in Liverpool bay where the offshore wind farm of West of Duddon Sands is located. The use of WRF compared to global NOAA GFS surface wind data to feed the WaveWatch III wave model can lead to a difference of up to factor two on the significant wave height. The resulting gain in scheduled operational working hours is significant and can be coarsely estimated as increasing by up to 12% in Liverpool Bay, 18% in Cardigan Bay or more than 20% in the Bristol Channel over a winter month. This increase could be even higher if the required duration of the weather window for carrying out specific marine operations were also factored in.

Some cost models for offshore wind generation [4], estimate that maintenance and operations account for around 20 to 33% of the total offshore project costs. A specific study shows that a wind farm with 100 turbines of 5MW, located 46km from the coast, would waste over €5.5 million due to weather downtime during a winter month [5]. The uncertainty forecasting the weather windows for the marine operations involved will have a significant impact in the final costs of operations and maintenance.

This study shows the importance of an adequate configuration of the wind and waves forecasting modelling tools employed for operational and maintenance purposes, suggesting significant economical benefits of this practise along the whole life cycle of the project.



Learning objectives
To understand the impact of the specific configuration of coupling two state-of-the-art models (WRF for the atmosphere and WW3 for waves) in the capability of forecasting relevant thresholds for operational planning for offshore wind farms.


References
[1] Tolman, H. L., 2002e: User manual and system documentation of WAVEWATCH III version 2.22. Tech. Note 222, NOAA/NWS/NCEP/MMAB,133 pp.
[2] Skamarock, W. C., J. B. Klemp, J. Dudhia, D. O. Gill, D. M. Barker, W. Wang, and J. G. Powers, 2005: A description of the Advanced Research WRF Version 2. NCAR Tech Notes-468+STR
[3] Cooper, W, Saulter, A, Hodgetts, P, Guidelines for the use of metocean data through the life cycle of a marine renewable energy development, 2008, CIRIA.
[4] B. B. A. Valpy, "How to Improve the Cost of Energy from Offshore Wind – Technology Perspectives," in RenewableUK 2010, Glasgow, UK, 2010.
[5] Maples, B.; Saur, G.; Hand, M.; van de Pieterman, R.; Obdam, T.(2013). Installation, Operation, and Maintenance Strategies to Reduce the Cost of Offshore Wind Energy.106 pp.; NREL Report No. TP-5000-57403.