Higgins P. and Foley A., 2013. Offshore Wind Power Forecasting Error and Electricity Market Implications. In: Proceedings of the 8th Conference on Sustainable Development of Energy, Water and Environment Systems (SDEWES2013), Dubrovnik, Croatia.

Offshore Wind Power Forecasting Error and Electricity Market Implications Higgins, P School of Mechanical and Aerospace Engineering Queen’s University Belfast, Belfast, Northern Ireland email: [email protected] Foley, A.M School of Mechanical and Aerospace Engineering Queen’s University Belfast, Belfast, Northern Ireland email: [email protected]

ABSTRACT Worldwide demand for wind power is increasing due to security of supply issues, increases in the price of fossil fuels and greenhouse gas emissions reduction targets. This increase in demand has resulted in a decrease in the availability of suitable onshore sites and as a result offshore wind power is growing rapidly. However, as wind is a variable resource and stochastic by nature, it is difficult to plan and schedule the power system under varying wind power generation. This difficulty is even greater for offshore wind power forecasting than onshore due to limited datasets and knowledge. This paper investigates the impacts of offshore wind forecasting error on the operation and management of the Irish pool-based electricity market. A Single Energy Market (SEM) model has been developed in PLEXOS for power systems to solve the unit commitment problem in 2025. The impact of different offshore wind forecast scenarios is analysed through the generation costs. The greatest impact of the offshore wind forecast error was on the import and export to the British market. KEYWORDS Offshore wind, electricity markets, forecast error, scheduling, dispatch INTRODUCTION The European Union (EU) has set a number of policy objectives [1] to achieve a sustainable energy future for Europe through measures to tackle climate change, to ensure energy security and to enhance competitiveness. Both the United Kingdom (UK) and the Republic of Ireland have high renewable electricity targets by 2020, 21% and 42.5% respectively [2, 3]. The Republic of Ireland and Northern Ireland have 2020 renewable electricity targets of 40%. In both countries renewable electricity comes mostly from wind power. Total installed wind power in the Republic of Ireland will be 4,000MW with 25MW offshore and 3,975MW onshore. In Northern Ireland the total installed wind power by 2020 will be 1,159MW with 191MW offshore and 968MW onshore [4, 5]. The all-island generating capacity for 2020 will be 15,636MW therefore the offshore wind in the Republic of Ireland and Northern Ireland make-up 0.2% and 1% of the generating capacity respectively. Offshore wind in the Republic of Ireland and Northern Ireland will have to be developed in greater quantities to achieve 2025 targets. Plans are underway to link and support this valuable renewable asset in the UK and Ireland. There has been the signing of a Memorandum of Understanding between the Irish and UK governments to export gigawatts (GW) of green energy from Ireland to the UK

Higgins P. and Foley A., 2013. Offshore Wind Power Forecasting Error and Electricity Market Implications. In: Proceedings of the 8th Conference on Sustainable Development of Energy, Water and Environment Systems (SDEWES2013), Dubrovnik, Croatia.

[6]. The disadvantage of wind power is that it is variable and stochastic by nature. The inclusion of wind power into the generation portfolio makes it difficult to schedule the power system under varying generation. Therefore wind power forecasting is a useful power system planning tool. Onshore wind power forecasting techniques have improved dramatically and continue to advance, but offshore wind power forecasting is more difficult due to limited datasets and knowledge. The offshore environment poses considerable problems for accurate wind power forecasting. It is smoother with fewer obstacles so wind speed and thermal effects are felt more acutely meaning wind variability is more prominent [7]. Therefore, as offshore wind generation increases the variability will become more pronounced and better planning techniques and forecasting will be required. Wind power forecasting tools are seen to be critical as they enable better dispatch, scheduling and unit commitment of thermal generators, hydro plant and energy storage plant and more competitive market trading as wind power ramps up and down [7]. As the amount of offshore wind power increases in the UK and Ireland more robust forecasting and planning techniques are required. This paper presents an investigate into the impacts of offshore wind forecasting on the operation and management of the 2025 single wholesale electricity market in the Republic of Ireland and Northern Ireland using PLEXOS for Power Systems. PLEXOS MODEL AND METHODOLOGY The SEM Market and PLEXOS From the early development of the Single Energy Market (SEM), both transmission system operators (TSO), EirGrid (Ireland) and the System Operator for Northern Ireland (SONI) selected Energy Exemplar’s PLEXOS for their SEM market software [8]. It is both robust and proven, and has been used to undertake a number of academic research studies of the SEM [9, 10, 11]. A study investigated wind power predictability as an investment factor for selecting onshore wind farm sites [12] found individual wind farms predictability as a poor design feature, as the reductions in imbalance cost are insignificant. However, the aggregation of individual wind farms was found to substantially increase the benefits of predictability. Along with improved financial benefits predictability can also play an important role in the operation and maintenance of offshore wind farms due to improved availability from reduced downtime periods. The impact of Ireland’s 2020 wind targets with varying limits for wind curtailment was examined and it was shown that approximately 7% to 14% of wind production could be lost [13]. The main factors in the reduction of wind curtailment were the wider spatial spread of wind farms and the higher overall capacity factors of offshore wind. The impact on the operation of the conventional generation during periods of high wind penetration was noted as significant. A study produced in 2007 investigated the total net benefits of grid integrated wind into the SEM [14]. Benefits were defined as capacity, emissions and fuel saving benefits. Capacity benefit was the saved cost of having to build and maintain another conventional generator instead of the wind generator. Emissions benefits are the reductions in carbon dioxide and

Higgins P. and Foley A., 2013. Offshore Wind Power Forecasting Error and Electricity Market Implications. In: Proceedings of the 8th Conference on Sustainable Development of Energy, Water and Environment Systems (SDEWES2013), Dubrovnik, Croatia.

sulphur dioxide emissions. The fuel savings are the savings from using wind rather than burning the fuel. The most significant wind generation benefits were achieved from increase interconnection between the SEM and BETTA, and a more flexible power plant portfolio with lower cycling units such as gas and less oil fired units. PLEXOS model Energy Exemplar’s PLEXOS for Power Systems version 6.208R03 was used with the Xpress optimiser [15]. Xpress was set with mixed integer programming at a relative gap of 0.05%. Xpress aims to minimise the objective function, shown in Equation (1) which is the conditional of a number of constraints: ∑

(1)

where t is the time period, cjt is the start cost of unit j in period t, Ujt is a binary quantity representing if unit j has started the period before t, njt is the no load cost of unit j, Vjt is a binary quantity representing the generating status of unit j, (1) is on or (0) for off, m jt is the production cost of unit j, Pjt is the power output of unit j, vl is the penalty for loss load, uset is the unserved energy of unit j and usrt is the reserve energy not met by unit j. The objective function is subjected to constraints such as maximum generation, minimum stable generation, minimum up and down times and ramp rates. PLEXOS requires generating/transmission parameters such as maximum and minimum ramp up/down time, minimum stable level, startup cost, forced outage rate, wind input data, operational and maintenance costs and fuel costs to calculate the objective function. These parameters are difficult to obtain normally as most generator datasets are confidential and proprietary. However, the Commission for Energy Regulation (CER) and Northern Ireland Authority or Utility Regulation (NIAUR) perform an annual validation exercise and release public available SEM datasets and software files [16]. These SEM .csv files contain the latest input parameter required to perform any modelling of the SEM. The short term schedule is selected as the focus of the work is over a year long period. The short term schedule is the most suitable as it is designed to emulate the dispatch and pricing of real market-clearing engines. The short term schedule optimises each of the 366 days in 2025 at 30 minute intervals. These settings should provide realistic results that represent the SEM dispatch scheduling. Test system The test model is based on the 2012 model developed and published by CER [16]. The 2012 model was modified to represent the all island market in 2025 (SEM_2025) with two generating units representing the BETTA market. The BETTA Gen is a single gas generating unit representing the entire conventional generating portfolio and the BETTA wind represents the wind generation in the BETTA. The BETTA wind is a time lag of the resource from wind region J. The BETTA wind generation was fixed with no forecast errors applied. The total installed 2025 generating capacities for the Republic of Ireland and Northern Ireland are 13,077MW and 4,624MW respectively. Two interconnectors are modelled between the SEM and the BETTA. The Moyle interconnector is from Northern Ireland to Scotland and the East West interconnector is from Ireland to Wales. The combined capacity of both connectors is 900MW. The demand forecasted is from the All-island Generation Capacity Statement for 2022 [5]. The medium growth projection for 2022 was linearly extrapolated to 2025. The 2025 demand forecast of 42,870GWh does not take into account future policy changes between 2022 and 2025. The all-island wind was modelled by separating the onshore wind

Higgins P. and Foley A., 2013. Offshore Wind Power Forecasting Error and Electricity Market Implications. In: Proceedings of the 8th Conference on Sustainable Development of Energy, Water and Environment Systems (SDEWES2013), Dubrovnik, Croatia.

into twelve regions and offshore wind into five regions [17], see Figure 1. Wind generation is scaled from 2008 records to achieve the 2025 targets. The 2025 total wind for the Republic of Ireland was 5,600MW with 4,600MW of onshore and 1,000MW of offshore. Northern Ireland has 1,890MW of wind with 1,290MW of onshore wind and 600MW of offshore wind. Research showed the offshore wind regions of the east coast of the Island have a one hour time lag to the respective onshore wind regions and the spatial correlation of neighbouring wind regions was found to be in the range of 0.94-0.97 [13]. The offshore wind data files were time and correlation adjusted accordingly to replicate this research.

Figure 1. SEM wind regions Forecasted fuel prices were obtained from [18]. Table 1 shows the fuel costs applied to the model. Currently the TSO has set the System Non Synchronous Penetration (SNSP) at 50%, this value is expected to rise to between 70% and 80% by 2020 [13]. For this 2025 model the SNSP is set at 75%. The TSO ensures the SNSP limit is not exceeded at any point throughout the year. The formula for SNSP is (2) Table 1. Fuel costs Fuel Type Coal Gas BETTA Gas Peat

Price (€/GJ) 2.12 7.02 7.5 3.18

SEM Design The SEM is a mandatory all-island wholesale pool market. The SEM is a pool through which generators and suppliers trade electricity. The market operates over three trading periods: before, intra-day and after. The before trading period consists of each generating unit bidding their commercial offer data (COD) and technical offer data (TOD) for each half hour interval in the intra-day. The COD and TOD contain price/quantity bids and no load costs. The price

Higgins P. and Foley A., 2013. Offshore Wind Power Forecasting Error and Electricity Market Implications. In: Proceedings of the 8th Conference on Sustainable Development of Energy, Water and Environment Systems (SDEWES2013), Dubrovnik, Croatia.

is the short-run costs which consist mainly of fuel related operating costs. SEMO schedules the optimised generation portfolio based on forecasted wind generation, the price/quantity pairs from all generating units and SEM system stability requirements. SEMO informs each generating unit of its required generation and operating hours before the start of intra-day trading. The intra-day period involves the real time application of the generation schedule. Both TSO’s implement the schedule and are responsible for delivering an efficient operation of the wholesale power market. The TSO’s continuously analysis the wind forecast and generating capacity requirements prior to the commencing half hour trading period based on updates to maintain system stability. SEMO calculates the System Marginal Price (SMP) over the four days after the intra-day. These four days are known as the after period. The SMP is calculated for each half hour trading period during the intra-day. This SMP is the price applicable to both generators and suppliers receiving/making payments for electricity generated/used [19]. Replicating SEM activities in PLEXOS The test system must mimic the SEM operation to fully investigate the impact offshore wind forecast error has on the wholesale market. Therefore a day ahead model is required to produce an optimised generation schedule similar to the operation of the before period in the SEM. This day-ahead generation schedule will feed into the real time model. The real time model will perform the activities of the intra-day period. The forecast error scenarios will be implemented in the real time model to determine the impact forecast error has on the operation and management of the SEM. The day-ahead model has a twenty four hour look ahead to replicate the SEM before period. Maintenance outages are included in the day-ahead model. The twenty four hour look ahead provides the model with enough time to develop an optimum schedule including the maintenance outages. The forecasted wind with no errors is included in this model. The key outputs from the day-ahead model are generation, undispatched capacity and short run marginal cost (SRMC). The day-ahead model creates .csv files for all three outputs containing half hour intervals values for the entire year. These .csv files represent the SEMO day-ahead schedules. The real time model uses the day-ahead .csv files and operates with a thirty minute look ahead. Maintenance and forced outages are applied to the real time model to replicate the realistic short notice plant outages. The wind forecast error scenarios are applied to the real time model to investigate the impact forecast error has on the market operation and management of the SEM. The after trading period of the SEM is not modelled for this research. Further research projects will involve the development of the after trading period. The impact of forecast error on the SMP will not be investigated in this paper but the generation cost will. Generation cost is calculated using the following formula ∑

(

(

)

(

))

(3)

where cri is the running cost. The running cost is an additional cost incurred for each hour that generating unit i is online. Offi is the fuel offtake, this is the fuel used during the start-up for the generator. Fi is the fuel price of the generating unit’s fuel. Pi is the power from unit i. VOM is the variable operation and maintenance charge. This is the charge used to recover the maintenance costs due to generation, such as wear and tear and other regular equipment replacement and servicing costs.

Higgins P. and Foley A., 2013. Offshore Wind Power Forecasting Error and Electricity Market Implications. In: Proceedings of the 8th Conference on Sustainable Development of Energy, Water and Environment Systems (SDEWES2013), Dubrovnik, Croatia.

WIND FORECAST ERROR EirGrid produce wind forecast accuracy statistics monthly. These reports provide information on the max forecast error when generation is more than forecast and when generation is less than forecasted. The report includes a Normalised Mean Absolute Error (NMAE). Eirgird perform this calculation as it puts the error in context of the total capacity. This method means the error can be compared to previous periods. The NMAE method is not applied to this research as the installed capacity for 2025 is assumed to be constant throughout the year. The following method was implemented to produce positive and negative scenarios to analyse the impact forecast errors have on the SEM. EirGrid publicly provide the wind generation and respective wind forecast data for every 15 minute interval from 2010 onwards [20]. This wind data is recorded from the SEM which consists of onshore wind generation. Offshore wind forecasted and generation data is difficult to source as it is not publicly available. It will be assumed that the calculated forecast error from the EirGrid data will be similar to the offshore wind forecast error. From this data the wind forecast error scenarios were determined using the following error difference formula. (4) where Wg is the generated wind (MW) and Wf is the wind forecast (MW). Figure 3 shows the distribution of the wind forecast error of wind generated in the SEM in 2012. The figure illustrates the frequency of each forecast error and which are the most dominate. A negative value, to the left of the centre represent less generation than forecasted. The columns to the right represent the error forecast when generation is greater than forecasted. It can be noted from both figures that the most significant wind forecast error occurs between +/-10% of the wind generation. The majority of forecast error occurs between +/-20% of wind generation.

Figure 3. Wind forecast error in the SEM

Higgins P. and Foley A., 2013. Offshore Wind Power Forecasting Error and Electricity Market Implications. In: Proceedings of the 8th Conference on Sustainable Development of Energy, Water and Environment Systems (SDEWES2013), Dubrovnik, Croatia.

RESULTS AND DISCUSSION A comparison of the results was carried out and the results are present in the following tables and figures. The impact of each scenario was compared over a range of different parameters, unit generation, unit operation time, unit start-ups and generation costs. Table 2 shows the generation comparison between all the scenarios in terms of GWh. The table represents the annual generation for each unit over the year, from the 1st of January 2025 to 31st of December 2025. The results reveals little change in the generation of coal, distillate oil, peat, waste, hydro and pumped hydro energy storage. The major generation changes are gas, BETTA generation and onshore wind. Figure 4 illustrates the alterations in generation for each scenario. At certain periods of the year the total generated wind is more than the SNSP limit and therefore both the onshore and offshore wind are curtailed. When the generated offshore wind is less than forecasted, more onshore wind can be accommodated before the SNSP limit is reached. The +/-20% extremes of the tests show onshore wind, BETTA generation and gas making up the difference for offshore wind. When the offshore wind generation is more than forecasted, onshore wind and gas are directed to curtail/reduce generation and less is imported from the BETTA. The Base model has 1,778 GWh of onshore wind curtailment due to the SNSP constraint. When the offshore wind forecast error is +20% an extra 733 GWh of offshore wind is added to the system. 175 GWh of onshore wind is further curtailed to accommodate this additional offshore wind. When generation is less than forecasted by -20%, 744 GWh of offshore wind is missing from the system so onshore wind curtailment is reduced. The wind curtailment does not account for all of the missing offshore wind, as only a reduction of 170 GWh to 1,608GWh is recorded. Table 2. Generation comparison across all scenarios. Scenario -20% -10% -5% -2% Base 2% 5% 10% 20% Generator GWh GWh GWh GWh GWh GWh GWh GWh GWh Gas 17,126 17,005 16,975 16,926 16,930 16,916 16,898 16,859 16,835 Coal 5,392 5,360 5,359 5,348 5,350 5,346 5,343 5,337 5,332 Distillate oil 3 1 1 1 1 1 2 1 1 Peat 2,264 2,252 2,248 2,245 2,245 2,244 2,242 2,239 2,237 Hydro 1,538 1,524 1,521 1,513 1,522 1,516 1,502 1,498 1,479 Pumped Hydro 359 363 360 362 362 360 365 365 375 Waste 395 394 393 391 390 390 389 389 386 Wind 14,779 14,706 14,636 14,625 14,610 14,584 14,558 14,522 14,435 Offshore wind 3,648 4,048 4,219 4,373 4,392 4,495 4,610 4,805 5,125 BETTA Gen 4,173 4,058 4,015 3,956 3,946 3,898 3,859 3,770 3,612 BETTA Wind 2,483 2,483 2,483 2,483 2,483 2,483 2,483 2,483 2,483

Higgins P. and Foley A., 2013. Offshore Wind Power Forecasting Error and Electricity Market Implications. In: Proceedings of the 8th Conference on Sustainable Development of Energy, Water and Environment Systems (SDEWES2013), Dubrovnik, Croatia.

Figure 4. Difference in generation across all scenarios Figure 5 illustrates the monthly exports from the SEM to the BETTA for the forecast error scenarios. The largest exports occur during the summer and autumn months. The SEM experiences a drop in demand from April to September, where it remains below 3,400GWh. The wind generation does not experience such a significant drop in generation. The seasonal drop in demand coupled with high wind generation results in an increase in exports from the SEM to the BETTA.

Figure 5. SEM exports

Higgins P. and Foley A., 2013. Offshore Wind Power Forecasting Error and Electricity Market Implications. In: Proceedings of the 8th Conference on Sustainable Development of Energy, Water and Environment Systems (SDEWES2013), Dubrovnik, Croatia.

The most operational time affected categories for the +20% scenario are the hydro and pumped hydro units. The decrease of 43MW in hydro generation resulted in a reduction of over 1,500 hours of operation over the year. The pumped hydro generation increase of 13MW resulted in an increase of over 500 hours of operation. The Hydro and pumped hydro are also the most started units in the SEM with almost 6,000 and 3,000 starts respectively. The next highest are the Gas and Wind generating unit with approximately 900 each. The hydro and pumped hydro are the most time affected and started units because they have some of the fastest ramp down/up capabilities of all the available generating units. When the generation is lower than forecasted the impact on the hydro operational hours is not as significant as gas units are switched on and more onshore wind is allowed onto the system. Figure 6 illustrates the impact on generation costs due to forecast error. The wind generation is modelled at a zero generation cost as it is a must run unit and has to be dispatched first. When the generation is greater than forecasted the generation costs are cheaper as more offshore wind is on the system. The Figure 6 (b) shows the difference in generation cost between the base and each scenario. Gas is the most affected while coal experiences a small increase when generation is less than forecasted. An important point to note is the significant difference in gas and coal generation cost when the generation is less than forecasted and when generation is greater than forecasted.

(a) (b) Figure 6. Generation costs across each error scenario Table 3 shows the impact of offshore wind forecast error on the annual generation costs if the 2011 and 2012 generation and forecasted profiles were applied to the 2025 model. The change in generation costs is greater when more generation than forecasted occurs. A greater change in annual generation cost is seen when generation is larger than forecasted for the +/2%, +/-5% and +/-10% error scenarios. The difference in annual generation costs for each scenario is multiplied by the percentage occurrence of the scenario to determine the effect of wind forecast error throughout 2025. The annual cost for each scenario is added and scaled to represent 100% of a full year. Table 3 shows the annual generation costs for 2011 and 2012 decrease by €1,681,000 and €1,144,000 respectively. These decreases represent less than 0.1% of the overall generation costs. If the BETTA generation was excluded from the costs and only the SEM generating units were analysed, an increase in generation costs would be noted.

Higgins P. and Foley A., 2013. Offshore Wind Power Forecasting Error and Electricity Market Implications. In: Proceedings of the 8th Conference on Sustainable Development of Energy, Water and Environment Systems (SDEWES2013), Dubrovnik, Croatia.

Table 3. Changes in generation cost for 2011 and 2012

Annual Generated cost (€000) -20% €28,960 -10% €12,426 -5% €7,360 -2% €276 2% -€4,113 5% -€7,782 10% -€16,345 20% -€27,782 Scenario Total Annual Total

Error Scenario

2011 % occurrence in the year 6% 7% 9% 9% 9% 9% 8% 7% 64% 100%

Weighted annual cost (€000) €1,797 €867 €634 €25 -€360 -€686 -€1,371 -€1,981 -€1,076 -€1,681

2012 % occurrence in the year 5% 6% 8% 8% 8% 7% 7% 6% 54% 100%

Weighted annual cost (€000) €1,585 €798 €556 €21 -€327 -€512 -€1,069 -€1,675 -€621 -€1,144

DISCUSSION AND CONCLUSION The study showed that offshore wind forecast error has the potential to have significant operation and management impacts on the SEM in terms of exports to the UK and total generation costs. The offshore wind capacity modelled for 2025 accounted for 9% of the total installed capacity. It was found that the offshore wind forecast error, for all of the scenarios, has a slightly greater impact on generation costs when the forecast is less than generated. Forecasting methods used by EirGrid provide a relatively normal distribution of the forecast errors. Combining the distribution and changes to yearly generation costs, the impact offshore wind forecast error has on the generation cost is not that significant. However, an increase to 2050 offshore wind targets could result in significant increases in generation costs. The preliminary research shows the BETTA generation is affected the most due to the offshore wind forecast error. A full model of the BETTA market is required to fully investigate the impact offshore wind forecast has on the SEM and the BETTA. The research has shown the significant impact the BETTA generation can have on generation, exports and generation costs. Further research involves developing the full BETTA market and analysing offshore wind on the combined BETTA and SEM for 2050. NOMENCLATURE BETTA CER COD EU GW GWh ISLES MW

British Electricity Trading and Transmission Arrangement Commissioner for Energy Regulation Commercial Offer Data European Union Giga Watt Giga Watt hours Irish-Scottish Links on Energy Study Mega Watt

Higgins P. and Foley A., 2013. Offshore Wind Power Forecasting Error and Electricity Market Implications. In: Proceedings of the 8th Conference on Sustainable Development of Energy, Water and Environment Systems (SDEWES2013), Dubrovnik, Croatia.

NIAUR NMAE PASA SEM SEMO SMP SNSP SONI SRMC TOD TSO UK VOM

Northern Ireland Authority for Utility Regulation Normalised Mean Absolute Error Project Assessment of System Adequacy Single Energy Market Single Electricity Market Operator System Marginal Price System Non Synchronous Penetration System Operator for Northern Ireland Short Run Marginal Cost Technical Offer Data Transmission System Operators United Kingdom Variable Operation and Maintenance

REFERENCES [1] European Commission, 2009. Directive 2009/28/EC of the European Parliament and the Council on the promotion of the use of energy from renewable sources. [2] Department of Energy & Climate Change. The UK Renewable Energy Strategy 2009. [3] Department of Communications, Energy and Natural Resources. Maximising Ireland's Energy Efficiency - The National Energy Efficiency Action Plan 2009 – 2020, 2009. [4] European Wind Energy Association, Pure Power – Wind energy targets for 2020 and 2030. 2011. [5] EirGrid and the System Operator for Northern Ireland (SONI), “All-Island Generation Capacity Statement 2012-2020. Generation Capacity Statement Report,” 2011. [6] Minister for Communications Energy & Natural Resources, Pat Rabbitte, TD, and the UK Secretary for Energy and Climate Change, Edward Davey. Memorandum of Understanding On cooperation in the energy sector, 24 January 2013. [7] Foley A.M, Leahy P.G, Marvuglia A, McKeogh E.J, “Current methods and advances in forecasting of wind power generation,” Renewable Energy, pp. 1-8, vol. 37, 2011. [8] Single Electricity Market Operator (SEMO), “Solver choice in the SEM – A comparative study of Lagrangian Relaxation vs. Mixed Integer Programming,” 2010. [9] Denny E., “The economics of tidal energy,” Energy Policy, pp. 1914-1924, vol. 37, 2009. [10] Deane J.P., Dalton G., O Gallachoir B.P., “Modelling the economic impacts of 500MW of wave power in Ireland,” Energy Policy, pp. 614-627, vol. 46, 2012. [11] Foley A., Tyther B., Calnan P., O Gallachoir B.P., “Impacts of Electric Vehicle charging under electricity market operations,” Applied Energy, pp. 93-102, vol. 101, 2012. [12] Girard R., Laquaine K., Kariniotakis G., “Assessment of wind power predictability as a decision factor in the investment phase of wind farms” Applied Energy, pp. 609-617, vol. 101, 2013. [13] McGarrigle E.V., Deane J.P., Leahy P.G., “How much wind energy will be curtailed on the 2020 Irish power system,” Renewable Energy, pp. 544-553, vol. 55, 2012. [14] Denny E., O’Malley M., “Quantifying the Total Net Benefits of Grid Integrated Wind” Power Systems, vol. 22, 2007. [15] FICO Xpress-Optimiser. Available at:< http://www.fico.com/en/Products/DMTools/xpress-overview/Pages/XpressOptimizer.aspx> [16] Commissioner for Energy Regulation. Available at: . [17] Redpoint Energy Ltd, “Single Electricity Market Operator (SEMO) forecast model of 2011-2012,” tech. rep., Commission for Energy Regulation (CER), 2011. [18] EirGrid, “Interconnection Economic Feasibility Report,” 2009. [19] SEMO, 2013. Available at:< http://www.sem-o.com/JoiningTheMarket/Pages/PricingandSchedulingFactsheet.aspx> [20] EirGrid, 2013. Available at:< http://www.eirgrid.com/>