Site Resource Characterisation and Assessment for Deployment of an Array of WaveSub WEC Devices.SummaryThe project is a part of the group proposal for the deployment of a 50MW array of Marine Power Systems’ WaveSub devices at one of the three sites: AMETS, Pembrokeshire Demonstration Zone and WaveHub, with consideration of the potential device power generation and relevant site wave resource. Some background knowledge on state of marine renewables industry as well as the device description and technical specification is provided for better understanding of this area. The chosen methodology for power calculations and utilization of wave roses is explained and justified, followed by demonstration and discussion of results – bivariate scatter diagrams, wave roses and omnidirectional power calculations. With other relevant factors in mind, a recommendation is given with regards to which site will be most appropriate for WaveSub array deployment in terms of available site wave resource. The overall results are made WaveSub device-specific as it incorporates assessment of sites followed by application of WaveSub power matrix.  Note: This report contains and utilises WaveSub power matrix, which is considered sensitive information. Parts of this report cannot be viewed, referenced and shared beyond persons, who signed the Non-Disclosure Agreement with Marine Power Systems.IntroductionIPCC released a report in 2018 in which almost a hundred scientists all around the world provided the proof of irreversible changes, caused by CO2 emissions. The changes include change in weather, that can be demonstrated by the increased temperatures during winters, rising sea levels, loss of certain ecosystems and their inhabitants, and increased levels of air pollution. In order to avoid the changes, the report calls for global warming reduction to 1.5°C (above pre-industrial levels)[1]. The report links back to climate strategies and emission targets, set out in different countries as well as in European Union, which outlines numerical goals, such as Greenhouse gas emissions, electricity and heat consumptions and share of renewable energy, compared to fossil fuel production[2]. Incidentally, the global warming reduction goals are currently at 2°C in the EU climate action plan. In order to achieve the set-out goals, multiple changes are needed, including social, economic and technological. A move towards solely green sustainable energy is required over the next few decades. Examples of renewables include solar, biomass, geothermal, wind and recently growing marine renewable sectors. Marine renewable technologies have had different extent of success over the past years. Although total offshore wind capacity is expected to reach just under 30 GW in 2020[3], total installed tidal capacity doesn’t exceed 10MW and wave energy converters seem to accumulate to under 1 MW total capacity[4], both capacities measured at the start of 2018. The main issue with wave energy converters (hereafter WECs) is technology. There seems to be no technology convergence for optimal design of a WEC[5]. EMEC defines some of the main types of WECs: Attenuator e.g. Wavenet; Submerged Pressure Differential e.g. WaveSub; Bulge Wave e.g. Pelamis, Anaconda[6]. However, more than 100 devices have been presented with tens of different operating principles with no superior type identified.Renewable wave energy technologies are still immature. Nevertheless, ocean and sea energy is endless; majority of WECs are submerged, therefore not affecting the seascape. Most WECs are based on simple physical principles and contain standard manufacturing components. The device discussed in this report is WaveSub, which operates based on pressure differential principle. A pressure differential is induced by sea waves rising and falling above the device, moving the top part, as shown in Fig. 1. The electricity is then generated by the fluid that is pumped by change in pressure[7].WaveSub has been in in development by Marine Power Systems (hereafter MPS) since 2008.Figure 1. Submerged Pressure Differential diagram. It illustrates the full motion that induces electricity. Credit: emec.org.ukProject DescriptionThis project consists of site selection and array design and assessment for deployment of a 50 MW array of WaveSub WECs on one of 3 proposed sites:WaveHub (England, near Hayle town)[8]Located 16km offshore from Hayle, with total area of 8km2 split equally among 4 berth and depths ranging from 51m – 57m.Installed export capacity – 30MW upgradable to 48MW.Pembrokeshire Demonstration Zone (Wales, near Pembroke Dock town)[9]Located 15-21kms off the South Pembrokeshire coast, with total area of 90 km2 and depths of 50-62m.Installed export capacity – 30 MW per connected device/array, maximum of 90 MW.AMETS (Ireland, near Belmullet town)[10], which consists of 2 sites:Belmullet Berth A, located  about 16km out from Belderra Strand, with total area of 6.9 km2 and average depth of 100m.Belmullet Berth B, located about 6km out from Belderra Strand, with total area of 1.5 km2 and average depth of 50m.Installed export capacity – 10 MW.Device specificationThe technical specification is provided by MPS via project briefing and subsequent emails.The device consists of 3 floats (see Fig. 2).Its physical dimensions are 120m x 35m.The rated power of a single device is 4.5 MW.Self-protection mechanism for conditions when significant wave heights exceed automatically lowers the device onto the bottom and ceases all WaveSub electricity generation.Minimum water depth for the device deployment is 50m with no max depth specified.Spacing between 2 devices is expected to be 5 reactor length all around the device, from the central point of each device as shown in Fig. 3.  R = 600m 35m120mFigure 2. Depiction of a WaveSub full scale device, its mooring systems and electrical connection cable. Picture is taken from MPS video presentation of the WaveSub device.Figure 3. The illustration of the spacing between WaveSub devices; not to scale. WaveSub drawing credit: Alice Walpole, project team member.Project management and organisationInitial recommendations on work load separation was provided in the project briefing as follows:Financial factorsSocietal supportLegal implicationsEngineering constraintsLogistical aspectsTechnological developmentThus, the project team of consultants chose their topics according to their area of expertise, with some amendments of the above suggested subject areas. This report contains resource characterisation and assessment aspect of the project, as it is a vital indicator of projects financial and technological success. In the latter case, the project testing will be incomplete if the array will be tested in very mild conditions, as opposed to the conditions it is designed for. In addition, storm protection technology must be tested, which is a key feature of WaveSub compared to other WECs.Gantt chart and meetings.In the first couple of weeks Nathan Collins and I produced a Gantt chart (Appendix 1), which we used to visualise key milestones and subject groups. It had very limited success due to inability to estimate everyone’s workload and couldn’t quantify the duration of research and calculations for their subject. That led to abandonment of the chart and introduction of progress presentation for each group meeting. Each consultant from our group would update everyone else about their findings and progress in their respective field prior to the start of each meeting. Compulsory meetings were held weekly on Monday, from 15:00 to 18:00; additional computer cluster meetings were held on demand on Friday mornings 10:00-12:00, which mostly the Engineering/Technological/Resource Assessment consultants used to work with software.MethodologyThis section will provide methodology which ensures consistency and accuracy in definition, estimation and analysis of the wave resource.The main aim of this section is to measure and describe the wave resource available for the device in each of the 3 proposed sites (by deriving site bivariate distributions of occurrences corresponding to the sea states defined by Hs and Tz), wave direction for array orientation (by presenting wave rose diagrams) and provide the estimation for power generated on each site (by combining the bivariate distribution and WaveSub power matrix, provided by MPS ).The suggestion for resource characterisation methods was provided by Equimar project. Dr Daniel Conley recommended looking into deliverables documents, specifically into Deliverable D2.2: “Wave and Tidal Resource Characterisation” [11]. It provides relevant guidance on producing bivariate distribution matrices and wave roses using software. The document was taken as a basis for this research. The software used for producing the diagrams is MATLAB.The required raw data was collected from wave buoys. The raw data required for above described analysis is as follows:Significant Wave Height (Hs)Mean Wave Period (Tz)Peak Wave DirectionIt should be noted, that for purpose of site assessment and comparison on fair and equal basis, year 2017 is chosen for acquiring data and calculating power. The fact that it is the previous year during the time this report is written provides reliability and sense of present data, which should represent more accurately the site resource in several years when the WaveSub array should be deployed. Such statement is made based on the comparison of 2005 historic deployment data[12] a few kilometres to the west from PDZ WaveRider buoy, when it was considered to represent current PDZ conditions. It showed to vary when compared to November 2018 PDZ WaveRider data and to the 2017 data that is used in this report for approximation of PDZ site conditions.Power CalculationsThe wave heights and mean periods can be used to calculate site’s energy flux (i.e. power). All 3 sites are described as deep water, because depths at individual points are more than half of corresponding wavelengths[13]. In deep water the wave power level is calculated_Pw=?*g264*?*Hs2*TzEquation 1. Omnidirectional wave power equation for deep water.Where,Pw  – Wave power (omnidirectional)?    – Water densityg    – Gravitational accelerationHowever, instead of approximating power generation for each site, it is possible to predict the exact value of power generated by a specific device in a given time period by creating a bivariate diagram and multiplying the frequency by device power at that specific sea state, provided in WaveSub power matrix. Therefore, the equation for evaluating total power output becomes_Pw=?(PWaveSubij*fij)Equation 2. Specific omnidirectional wave power output equation for WaveSub.Where,Pw           – Wave power (omnidirectional)PWaveSub – Value for power from WaveSub power matrixf              – Frequency of specific sea state occurrence from site bivariate diagram (%)i,j              – row and column numbers for a given sea state.Wave directionWave direction and corresponding wave heights are represented by wave rose diagrams, which serve multiple purposes. Firstly, it can be used to calculate directional power, as some WECs conversion efficiency will depend on how unidirectional waves are. Waves coming from adjacent directions to the main one would only generate a fraction of power compared to main direction waves. However, this requires knowledge of how much WaveSub device power generation falls off depending on incoming angular waves, which was not provided. Thus, only omnidirectional power is used in calculations and resource assessment. Wave direction is fundamentally required to align the array to most incoming waves to achieve maximum power generation, thus essential for array design. Secondly, wave direction is necessary for coastal process impact assessments, which is discussed by Enno van der Linde as a part of this project.Wave rose diagrams were available for all sites apart from PDZ, for which it was manually created.Data quality and potential sources of errorThe three types of raw data discussed earlier are obtained from wave buoys. The data quality and availability, however, is not the same for each site.WaveHubPDZAMETS BAMETS AName of the buoyWaveHubSeacams2 PDZBelmullet BBelmullet AType of buoyWaveRiderWaveRiderWaveRiderWaveRiderData availabilityWhole year, every 30 min. Downloaded from: www.channelcoast.orgPast 2 days, every 20 min. Downloaded from: www.cefas.co.ukBuoy was active from 05/02/2018 until 18/11/2018.Whole year, every 30 min, Downloaded from: oceanenergyireland.ieDepth of the buoy (m)5255~50~100Location of the buoy*[14][15][16]Table 1. The illustration and comparison of buoy type and buoy data available for each site.*Pictures are available in full size in Appendix 2Note: The location is indicative of how accurate the buoy represents the average data of the site. In all three cases the buoy is deployed at depth greater than 50m (see Table 1 above), which indicates appropriate data for evaluation of sites for WaveSub, which can’t be deployed shallower than 50 m sites.Lack of data at PDZ and estimation methodologyThe PDZ wave buoy had insufficient data for power calculations and resource assessment. This is because:The wave buoy did not collect data for a full year and had no data for previous years.The wave buoy did not collect any winter seasonal data, although that is the season when WECs tend to produce highest power output due to highest waves and longest wave periods.The wave buoy data collection was ceased on 18 November 2018.Therefore, an estimation method was required to assess suitability of the site in terms of energy resource. The wave direction from PDZ wave buoy suggested clear majority (>97%) of waves coming from southwest, as will be presented in the results section of this report. Therefore, a collection of data from buoys upstream and downstream was considered. A buoy 45 km directly downstream from PDZ wave buoy was chosen as it matched the criteria, as shown in Fig. 4. This buoy is deployed at Scarweather Wavenet site and has a whole year of data for every 30 minutes[17], identical to type of data found for WaveHub and AMETS.45 kmFigure 4. Depiction of relative locations of PDZ WaveRider and WaveNet Scarweather buoys. Obtained from www.cefas.co.uk.After comparison of wave heights and wave periods, an amplitude vs offset method was chosen. It is assumed that waves from PDZ buoy travel towards east continuously, losing in wave height and wave period amplitudes. So, based on average Hs and Tz for each site during 24-hour period between 16-17 November 2018 an offset for height and another offset for amplitude was added to WaveNet buoy values to simulate PDZ WaveRider. The result was visualised and assessed on credibility as Appendix 3 depicts.The estimation, therefore, was considered appropriate and every WaveNet buoy Hs value was corrected by “+1.192” and Tz value by “+1.462” (difference between the original buoy data averages). This leaves a relatively large margin for error, as the average difference for an entire year is estimated based on 24-hour data. This is a case where data is missing and will require further observations to gain a full range of reliable data in order to improve accuracy of the site resource assessment.Results and DiscussionBivariate DiagramsA set of bivariate energy diagrams created for WaveHub as an example is provided below in Fig. 5 and Fig. 6, but all individual diagrams for each site are showcased in Appendix 4 for comparison. They are histogram-type diagrams that display % occurrence of particular sea states. As an example, the WaveHub diagram illustrates that the most common sea state is Hs = 1-1.5 m and Tz = 4-5, which occurs in 9.53% out of all present waves. Darker colours indicate higher % occurrence. Wave power curves were provided to indicate the wave resource available at certain sea states. Note: power curves do not extend further than Hs=10m, as anything above is considered to be storm conditions at which WaveSub will produce no power.Figure 5 Bivariate annual energy scatter diagram for WaveHub site. Created in MATLAB using WaveHub buoy wave data.Figure 6 Bivariate winter seasonal energy scatter diagram for WaveHub site. Created in MATLAB using WaveHub buoy wave data.From the comparison of Fig.5 and Fig.6 it can be inferred that winter period provides better wave resource than annual average value. Therefore, winter season diagram is often produced alongside annual diagram to demonstrate how effectively power will be generated during the best wave climate season for WECs, expressed as Capacity Factor.In addition, a general trend can be seen from the diagrams that more power can be produced at sea states, which approach right upper corner, increasing Tz and Hs. This tendency is confirmed by wave power equation that shows that wave power is directionally proportional to Hs2 and Tz (eq. 1). Likewise, comparison of the 4 site bivariate diagrams in Appendix 4 suggests AMETS sites to have greater wave resource annually and seasonally due to higher occurrence of waves with large wave height amplitudes and longer wave periods. The bivariate scatter diagrams provide the % occurrence ( fij) for individual sea state cells for final power calculations.Power MatrixPower matrix is sensitive information provided by MPS under individually signed Non-Disclosure Agreement (NDA) terms of every member of this project. It represents instantaneous power that one WaveSub device can generate at a certain sea state.Figure 7. WaveSub Power Matrix, provided by MPS. Number in each cell is power (in kW) that the device can produce at that sea state.It can be suggested why bivariate diagrams alone can’t be used to represent available wave resource for a WEC device: the earlier diagrams indicated just the trend of increase in power with increase in Hs and Tz, whereas the power matrix (Fig.7) indicates that power eventually falls off with long wave periods. In addition, the survivability mechanism means that the power during Hs above 10m will not be produced as also seen in the matrix above. This table provides PWaveSubij values for Eq.2 calculations. In addition, the matrix defines device rated power as 4.5 MW.Final Power calculationWith all variables in Eq.2 defined, it can be used for final power calculations, which will rank the sites from lowest wave resource to highest.WaveHubPDZAMETS BAMETS APw for one WaveSub device would be deployed at the site in 2017 – annual (MW)0.941.251.691.95Pw for one WaveSub device would be deployed at the site in 2017- Winter seasonal (MW)1.291.322.473.18Capacity Factor – annual (Pw/Prated)0.210.280.380.43Capacity Factor – Winter seasonal (Pw/Prated)0.290.290.550.71Total annual power would be generated in 2017 (GWh)8.1910.9514.8017.08Table 2. WaveSub device power calculations.The results from the Table 2 are mostly representative of the site resource bivariate diagram trend. AMETS had the highest resource, whilst WaveHub showed the least potential. Similarly, if WaveSub would be deployed in AMETS A it would generate double electricity of the number if it would be deployed at WaveHub. The annual power generation gives a clear indication of the site ranking:AMETS AAMETS BPDZWaveHubLikewise, the Capacity Factor is the highest for AMETS A, indicating high efficiency of the device converting energy on the site. Usually due to calm sea conditions during Summer period, Winter seasonal Capacity Factor is taken to represent the device. Taking in comparison a similar, 5.9MW capacity device (but with a different electricity generation principle), Wave Dragon can work at 0.43 of its full capacity [i.e. Capacity Factor = 0.43]. On the other hand, a technology which resembles WaveSub principles closer – Platoon Power Converter(PPC), operating as a point absorber, can only achieve 0.095 Capacity Factor. PPC has capacity of just above 3.6 MW[18]. Therefore, all sites considered in this report provide a considerable Capacity Factor for the device, with AMETS A showing outstanding value of 0.71 for Capacity Factor.The power calculations above provide omnidirectional power output for the device, whereas the direction of the waves and the spread of directions will affect the true value for power output.Wave rose diagramsAs discussed in methodology, these diagrams provide visualisation of wave climate at each site and how close to unidirectional waves are. All diagrams are available in full size in Appendix 5.                            WaveHub                     PDZ                           AMETS B                              AMETS AFigure 8. Wave Rose diagrams for each site. The length of each cone represents the occurrence of that wave direction, compared to other wave directions present on the spectrum. References are provided in Appendix 5.From the wave rose diagrams (Fig. 8) it can be suggested that depending on WaveSub sensitivity to different wave directions, AMETS sites power output will be affected the most by the spread in wave directions. Similar scenario will be applicable to WaveHub to some extent. Conversely, PDZ omnidirectional power should be almost identical with the true power output due to virtually no spread in incoming wave directions.Conclusions and further recommendationsRelevant information from project team membersAMETS B can currently fit only 2 devices due to area restrictionsAMETS A and B can currently fit only 2 devices due to extraction capacity limitation to 10 MW cable.ConclusionsOmnidirectional annual power output indicates an advantage of deploying a WaveSub device at AMETS sites, with preference for AMETS A. Winter season Capacity Factors are outstanding for AMETS but comparison to other devices suggests that WaveHub and PDZ Capacity Factors can be considered to be average or even above average for WECs. Due to immaturity of wave renewable energy industry it is difficult to assess Capacity Factors, as there is no same principle commercial device similar to WaveSub that can be used to draw a comparison between devices. Furthermore, it is difficult to assess the impact of wave direction spread on true power generated, so these results and conclusions rely heavily on omnidirectional power. From wave resource perspective, AMETS shows the best results. However, considering an array of 10-11 devices will be required to achieve around 50 MW of power total, it is currently impossible to use AMETS to achieve 50 MW. Therefore, only PDZ and WaveHub are currently viable if no changes will be made to electrical infrastructure for both AMETS sites and site expansion for AMETS B.RecommendationsTherefore, recommendations, based on the results of site resource assessment and assessment of other factors by other consultants on this project, are presented below:Currently with the given site and electrical infrastructure conditions, PDZ and WaveHub are the only options.Due to multiple approximations for PDZ site and very similar winter capacity factor, WaveHub would be the recommended site.If annual wave data for PDZ site will be acquired from SeaCams2 WaveRider buoy or otherwise, both sites will need to be compared for wave resource again to make a choice between the two sites.If electrical infrastructure will be upgraded in AMETS from 10 MW to 50MW+, AMETS A will be the preferred choice for WaveSub array deployment in terms of wave resource. Further workAcquire wave data for PDZ site.It is recommended to undertake assessment of how wave directional spread affects power generated by the WaveSub device.Effect of this report on the group report and group presentationArray design and coastal processes impact assessment will be based on some results of this report.AppendicesAppendix 1Gantt Chart, developed By Nathan Collins and Philipp Shitarev. Its purpose is tracking everyone’s progress with regards to each project team member.Appendix 2Location of wave buoys with regards to site perimeters. It illustrates how accurate wave buoy data represents site average data.WaveHub buoyPDZ buoyAMETS buoysAppendix 3Hs and Tz comparison between PDZ and Wavenet scaled buoys. There is little difference observed between the amplitudes in both estimations.Appendix 4Site specific bivariate annual and winter diagrams, produced in MATLAB and used for power calculations.WaveHubPDZAMETSAppendix 5 Wave Rose diagrams for each site. For AMETS and WaveHub the rose diagrams were downloaded from relevant reports, whereas PDZ was drawn using wave direction data gathered by Scarweather WaveNet buoy.Downloaded from www.channelcoast.org.Produced in Excel using Scarweather WaveNet directional data.Both wave rose diagrams are taken from oceanenergyireland.ie