OPTIMUM MEDIUM SCALE WIND-CAES CONFIGURATIONS FOR THE ELECTRIFICATION OF REMOTE COMMUNITIES

Transcription

1 OPTIMUM MEDIUM SCALE WIND-CAES CONFIGURATIONS FOR THE ELECTRIFICATION OF REMOTE COMMUNITIES Achilleas Marcogiannakis Petros Pasas Dimitrios Zafirakis 1 John K. Kaldellis Lab of Soft Energy Applications & Environmental Protection, TEI of Piraeus, Greece, P.O. Box 4146, Athens 1221, Greece, jkald@teipir.gr 1 Norwich Business School, University of East Anglia, Norwich, NR4 7TJ, United Kingdom ABSTRACT Although renewable energy sources (RES) technologies are nowadays granted as established, back up power is still required in order to support the variable energy generation owed to e.g. the stochastic nature of wind speed. In this context, there are various energy storage technologies that may interact with the primary RES energy source and achieve 1% energy autonomy for the load consumption each time investigated, eliminating at the same time contribution of thermal power generation. Among the various energy storage solutions, grown interest is recently noted in the investigation of compressed air energy storage (CAES) systems. Acknowledging the fact that although CAES is considered as bulk energy storage, it may equally well be downscaled so as to serve small-medium size applications, the current research study investigates the solution of an integrated Wind-CAES solution used to serve remote communities. More precisely, an optimization methodology is developed on the basis of a techno-economic analysis under the restriction of 1% energy autonomy achieved. Furthermore, several areas of different wind potential quality are investigated, while on top of that two different system versions are studied; i.e. the conventional CAES cycle and the dual-mode CAES cycle, where the system may allow shift to the Brayton cycle when energy stores are not sufficient to cover energy demand. Results obtained indicate the feasibility of the proposed solution, while the importance of exploiting a high wind potential is reflected. Finally, the advantage of the dual-mode CAES solution over the conventional CAES cycle for areas of low quality wind potential is designated. 1. INTRODUCTION Increased interest is recently noted in the promotion of the so-called distributed generation (DG) [1], with renewable energy sources (RES) technologies called to play a critical role in the shift attempted from centralized power generation to DG patterns. At the same time, there are several areas across the globe that cannot appreciate connection to a solid electricity grid and thus rely on stand-alone energy solutions, normally employing autonomous power stations operating on imported oil quantities [2]. On the other hand, in many of these regions one may encounter medium to high quality RES potential that encourages installation of solutions such as wind power and photovoltaics. Although such technologies are nowadays granted as established, back up power is still required in order to support the variable energy generation owed to e.g. the stochastic nature of wind speed, and thus satisfy energy demand at all times. In this context, there are various energy storage technologies [3], either more mature or emerging, that may interact with the primary RES energy source and achieve 1% energy autonomy for the load consumption each time investigated, eliminating at the same time the contribution of thermal power generation. Among the various energy storage solutions existing, grown interest is recently noted in the investigation of compressed air energy storage (CAES) systems, normally used in energy management applications. Operation of such systems is based on the exploitation of waste/surplus (e.g. wind energy curtailments) or off-peak low price energy amounts, feeding a motor-compressor system that compresses air under high pressure at either an underground air cavern or a high pressure tank. When an incentive to sell energy (i.e. during peak hours) or an energy deficit (i.e. when demand is high and RES energy production is not sufficient) appears, air is drawn from the cavern/tank and mixes with natural gas to produce high enthalpy gases in a combustion chamber, then used to operate a gas-turbine for power generation. The benefit arising from system operation is twofold; exploitation of otherwise waste or cheap energy on the one hand and considerable reduction of fuel on the other 1

2 (in comparison to the conventional gas-turbine (Brayton) cycle, where the net power output is reduced owed to the fact that almost 2/3 of the gas-turbine output is used to operate the compressor). Acknowledging the benefits arising from system operation and the fact that although CAES is considered as bulk energy storage, it may equally well be downscaled so as to serve small-medium size applications, the current research study investigates the solution of an integrated Wind-CAES scheme used to serve remote communities. More precisely, an optimization methodology is developed on the basis of a techno-economic analysis under the restriction of 1% energy autonomy offered to the remote community each time examined. Furthermore, to examine the effect of the local wind potential on the results obtained, several areas of different wind potential quality are investigated, while on top of that, two different system versions are studied; i.e. the conventional CAES cycle and the dual-mode CAES cycle, where the system may allow shift to the Brayton cycle when energy stores are not sufficient to cover energy demand. Results obtained indicate the feasibility of the proposed solution, while the importance of exploiting a high wind potential is reflected. Finally, the advantage of the dual-mode CAES solution over the conventional CAES cycle for areas of low quality wind potential is designated. 2. DESCRIPTION OF THE CAES SOLUTION In a CAES system, off-peak power is taken from the grid or other generation sources and is used to pressurize air into an underground cavern (pressures even reaching 8bars). During times of peak demand, the required amount of air is released from the cavern, heated with natural gas and then supplied in the form of gases to a gas-turbine where expansion takes place as in a typical Brayton/Joule cycle. And this is actually the main benefit of a CAES system, i.e. the fact that the stages of compression and generation are separated from one another. Consequently, what seems to be as much as 66% of fuel consumption for the compressor to be driven in a typical Brayton/Joule cycle, is not the case for the CAES cycle. As a result, in a CAES system, the entire power of the gas-turbine is available to the consumption. In this context, during a charging/discharging cycle, 1kWh of generated electricity requires approximately.75kwh of electricity for the compressor and 4,5kJ of fuel for the air combustion [4]. This required amount of fuel is the main subject of controversy over the unconditional acceptance of such systems. In an effort to disengage the CAES from the natural gas factor, one concept supports the use of biofuel [5], while another interesting approach is the so called "Advanced Adiabatic CAES" where no fuel is used [6]. On the other hand, CAES requires favourable sites and geological formations, suitable for underground storage. The storage media most commonly used are rock caverns, salt caverns, porous media reservoirs and buried pipes for small subsurface CAES units [7]. In terms of energy capacity, CAES is thought to be the only reliable alternative option to pumped hydro storage. Since the losses recognized are not appreciable, storage period is considerable, while among the advantages of CAES one may also include the fast ramp rates (2 to 3 times faster than conventional units), the stable heat rate at low capacity, and the considerably lower emissions (compared to simple and combined cycle units). Note that because of their potential to operate at partial load with satisfying fuel consumption, CAES systems are more suitable for load control applications. Besides, the flexibility of CAES systems to serve as both base load plants [5,8] and peak following units [9] provides considerable opportunities for improved management of wind energy generation [1,11]. 3. DESCRIPTION OF THE DUAL-MODE SOLUTION Emphasizing on DG energy patterns and the need to achieve energy autonomy using a combined Wind-CAES scheme, an alternative solution is currently proposed. More precisely, to counterbalance the need for extreme wind power and energy storage capacity so as to achieve 1% energy autonomy of a given remote area, the solution of a dual-mode CAES plant is currently adopted. Such a plant has the ability to switch its operation from the CAES mode to the traditional gas-turbine cycle with the help of a clutch that may allow connection between the gas-turbine and the compressor. To this end, the proposed solution (see also Fig. 1) comprises of the following main components: A wind park comprising of a number of wind turbines with total capacity "N WP ". A motor of "N m " rated power, used to exploit any wind energy surplus and feed the compressor under an efficiency of "η m ". A multi-stage compressor, used to compress ambient air into the air cavern/tank, under a given pressure ratio "r c ". Similar to the case of the motor, the compressor power "N c " is currently configured in relation to the maximum possible energy surplus appearing, i.e. "N WP -N min ", where "N min " is the minimum energy consumption encountered in the system examined, taking also into account any energy losses induced by the motor. A storage cavern or tank of maximum volume storage "V ss " and maximum depth of discharge "DOD L ", determined by the ratio of [(r c -r t ) r c -1 ], where "r t " is the pressure ratio of the gas-turbine employed. A preheater of efficiency "η pr ", used to preheat the air released from the cavern. A combustion chamber where the required amount of 2

3 compressed air and natural gas are mixed together for the production of gases that will operate the gasturbine, under a temperature of "T cc ". A natural gas tank, used for the fuel storage, with the latter being determined by the respective calorific value (CV) "H u ". A gas-turbine of a certain power output "N t ", directly related with the maximum appearing energy deficit, i.e. N t N def-m η t-g -1, with "η t-g " being the total efficiency of the gas-turbine and electricity generator. Fig. 2: Screen-shot of the Wind-CAES-DM algorithm In this context, the operation scenarios of the proposed configuration include the following: Fig. 1: The proposed dual-mode Wind-CAES system In this regard, the main problem variables currently taken into account include the wind farm capacity and storage volume, while main problem inputs require detailed wind speed and ambient temperature-pressure measurements along with the hourly electricity load demand of the system under investigation. At the same time, the technical characteristics of main system components are also required (see also Table 1 and [12] for more information on the developed sizing algorithm), while to simulate operation of similar systems, a sizing algorithm has been developed in Visual-Basic (see also Fig. 2). TABLE 1: MAIN ENERGY PROBLEM INPUTS Motor power "N m " Function of N WP ;N min Motor efficiency "η m ".9 Compressor power "N c " (MW) Function of N c ;η m Compressor eff. "η c ".8 Storage volume "V ss " (m 3 ) Variable Preheater eff. "η pr ".5 Gas-turbine power "N t " (MW) Function of N def-m ;η t-g Gas-turbine-gener. eff. "η t-g ".8 Gas-turbine cycle eff. "η t-g-gt ".35 Air to fuel ratio "λ" 45 Natural gas CV "H u " (MJ/kg) 47 Ambient temp. "T amb " (K) Time series Cavern temp. "T ss " (K) 298 Ambient pressure "P amb " (bar) 1 Max compression ratio "r c " 7 Min expansion ratio "r t " 3 Air heat capac. "C p " (kj/kg.k) 1.4 Gas heat capac. "C g " (kj/kg.k) 1.1 Combustion temp. "T cc " (K) 1256 A. In the case that wind energy production is sufficient to cover energy demand, wind energy is directly fed to the local consumption and any appearing energy surplus is used to compress air inside the cavern, provided that the latter is not full. B. In the case that wind energy production is not sufficient to cover the load demand, the required amount of compressed air and fuel are drawn in order to operate the gas-turbine. C. In case that both wind energy and energy stores are not able to cover load demand, the appearing energy deficit is covered by the dual-mode system operation, i.e. the gas-turbine is used to operate the compressor and produce the appropriate energy, under a different heat rate or efficiency "η t-g-gt " in comparison to the CAES cycle. As a result, given a wind farm capacity value, the hours of load rejection per year are recorded under a fixed storage volume, while to obtain minimum hours of rejection the storage capacity is gradually increased within a predefined range of variation. Furthermore, in the case that energy autonomy is not achieved, the wind park capacity is also increased, up to the point that 1% energy autonomy is obtained on the basis of using the Wind-CAES solution. At the same time however, results obtained also include the complementary energy required by the dual-mode CAES cycle in case that 1% energy autonomy is not achieved by the original Wind-CAES system. 4. APPLICATION TO REPRESENTATIVE WIND POTENTIAL CASE STUDIES The proposed solution is accordingly applied to three different wind potential areas, being representative of a low, medium and high wind potential case of islands 3

4 found in the Aegean Sea, Greece. Note that the respective regions correspond to isolated electricity systems, depending heavily on oil imports, while at the same time appreciating in most cases respectable RES potential, including also wind energy. In the meantime, plans concerning the introduction of LNG terminals in certain island areas stimulate investigation of the Wind-CAES solution. In this context, annual wind energy measurements of the three representative areas currently used are given in Fig. 3, with the annual mean wind speed at 1m height corresponding to 8.2m/sec, 6.2m/sec and 4.7m/sec respectively. Wind Speed (m/sec) Annual Wind Speed Measurements on an Hourly Basis for the Three Areas of Investigation High Wind Medium Wind Low Wind Hour of the Year Fig. 3: Annual wind potential of areas examined Meanwhile, the hourly load demand profile of a medium scale island for an entire year is given in Fig. 4, with the peak load demand reaching 6MW and the respective minimum load demand dropping to 1MW, while the annual energy demand exceeds 3GWh. Furthermore, a typical wind turbine power curve currently used in order to estimate wind energy production on the basis of wind potential measurements available is shown in Fig. 5. Load Demand (MW) Annual Variation of Load Demand on an Hourly Basis (Medium Scale Island) Hour of the Year Fig. 4: Annual load demand of a representative island area Next, by applying the proposed methodology, results obtained by the algorithm are presented in the following figures, considering a detailed energy balance analysis on an hourly basis for the entire year. More precisely, in Figs. 6-8 one may obtain the results obtained for all three areas investigated, concerning hours of load rejection per year for the operation of the Wind-CAES scheme Non Dimensional Power Output 1,2 1,,8,6,4,2 Non Dimensional Power Curve of a Typical Wind Turbine, 2,5 5 7,5 1 12, ,5 2 22,5 25 Wind Speed (m/sec) Fig. 5: Typical wind turbine power curve Hours of Load Rejection per Year the Levels of Energy Autonomy (Low Wind Potential Case) V=1,m3 V=8,m3 V=6,m3 V=4,m3 V=2,m3 V=9,m3 V=7,m3 V=5,m3 V=3,m3 V=1,m Fig. 6: Energy autonomy results (low wind case) Hours of Load Rejection per Year the Levels of Energy Autonomy (Medium Wind Potential Case) V=1,m3 V=8,m3 V=6,m3 V=4,m3 V=2,m3 V=9,m3 V=7,m3 V=5,m3 V=3,m3 V=1,m Fig. 7: Energy autonomy results (medium wind case) Hours of Load Rejection per Year the Levels of Energy Autonomy (High Wind Potential Case) V=1,m3 V=8,m3 V=6,m3 V=4,m3 V=2,m3 V=9,m3 V=7,m3 V=5,m3 V=3,m3 V=1,m Fig. 8: Energy autonomy results (high wind case) 4

5 In this regard, variation of both main parameters, i.e. wind power capacity and storage volume, is examined within the predefined ranges of 4-6MW and 1,-1,m 3 respectively. According to the results provided, increase of wind power capacity gradually increases the levels of energy autonomy achieved, with the parallel increase of the storage volume allowing for greater exploitation of the resulting wind energy surplus, thus leading to the reduction of load rejections per year. In fact, as one may see, energy autonomous configurations (i.e. configurations that guarantee zero load rejections for the entire year period) are also designated in all cases examined. To this end, in the case of low wind energy potential, one needs a wind farm capacity that exceeds 5MW and a storage volume in the order of 1,m 3. On the other hand, in the case of the medium wind potential, energy autonomous configurations result for wind power capacity that is higher than 4MW, with the respective minimum storage capacity required even approaching 5,m 3 for the highest wind power capacity, i.e. 6MW (half the one corresponding to the low wind potential case). Finally, in case that a high wind potential area is taken into account, wind farm capacity of even 14MW is able to provide 1% energy autonomy, if the highest storage capacity can be employed. Furthermore, if excluding the case of using storage capacity in the order of 1,m 3, all other storage capacity values examined may guarantee zero load rejections throughout the year, provided of course that a minimum wind power capacity is used. At the same time, the algorithm may also provide the long-term storage level variation of the air cavern; see for example Fig. 9, where mass of stored air ranges between the respective maximum "M ss " and minimum (1- DOD L ) M ss values, reflecting at the same time the ability of the high wind potential area to restrict -almost entirelydepletion of the air cavern energy stores for the respective example (see also Fig. 8). Mass of Air (x1 6 kg) 4,2 3,7 3,2 2,7 2,2 1,7 4 Variation of Storage Levels for the Three Different Wind Potential Cases (V ss =5,m 3, N WP =12MW) Hour of the Year Low Wind Potential Medium Wind Potential High Wind Potential Fig. 9: Variation of storage cavern air mass levels for different wind potential areas Accordingly, the algorithm also calculates the annual fuel consumption attributed to the operation of the CAES cycle only (see also Figs. 1-12). As one may see, there is a vast increase of CAES fuel consumption for the early stages of wind power capacity increase, while maximum fuel consumption is recorded once energy autonomy levels approximate 1%. CAES Fuel Consumption (t NG) CAES Fuel Consumption (Low Wind Potential Case) V=1,m3 V=8,m3 V=6,m3 V=4,m3 V=2,m3 V=9,m3 V=7,m3 V=5,m3 V=3,m3 V=1,m Fig. 1: CAES fuel consumption results (low wind case) CAES Fuel Consumption (t NG) CAES Fuel Consumption (Medium Wind Potential Case) V=1,m3 V=8,m3 V=6,m3 V=4,m3 V=2,m3 V=9,m3 V=7,m3 V=5,m3 V=3,m3 V=1,m Fig. 11: CAES fuel consumption results (medium wind case) CAES Fuel Consumption (t NG) CAES Fuel Consumption (High Wind Potential Case) V=1,m3 V=8,m3 V=6,m3 V=4,m3 V=2,m3 V=9,m3 V=7,m3 V=5,m3 V=3,m3 V=1,m Fig. 12: CAES fuel consumption results (high wind case) From that point onward, fuel consumption is reduced due to the increased participation of wind energy, opposite to the maximum CAES participation recorded for the minimum wind power capacity ensuring load rejections near zero. This is more clearly obtained in the case of the high wind potential case, where maximum CAES fuel consumption is noted between 12-16MW for storage volumes exceeding 1,m 3. 5

6 Moreover, as one can see, the impact of the local wind potential is of primary importance, with the required fuel amount even exceeding 16 tonnes of natural gas for the low wind potential and the highest storage volume achieving full energy autonomy. The respective value for the medium wind potential area reduces to marginally exceed 12 tonnes, while in the case of the high wind potential area, minimum wind farm capacity achieving energy autonomy implies CAES fuel consumption below 1 tonnes per year. On the other hand, if allowing the wind farm capacity to increase even further, CAES contribution reduces remarkably, even dropping below 75 tonnes of natural gas per year. Finally, results obtained by the algorithm execution also include calculation of fuel consumption attributed to the complementary operation of the dual-mode CAES cycle, i.e. the typical gas-turbine cycle (see also Figs ). In this regard, the option of using zero storage capacity is also examined, actually corresponding to the parallel operation of the wind farm and a typical gas-turbine plant. Dual Mode Fuel Consumption (t NG) Dual Mode Fuel Consumption (Low Wind Potential Case) V=1,m3 V=9,m3 V=8,m3 V=7,m3 V=6,m3 V=5,m3 V=4,m3 V=3,m3 V=2,m3 V=1,m3 Zero Storage Fig. 13: Dual-mode cycle fuel consumption (low wind case) Dual Mode Fuel Consumption (t NG) Dual Mode Fuel Consumption (Medium Wind Potential Case) V=1,m3 V=9,m3 V=8,m3 V=7,m3 V=6,m3 V=5,m3 V=4,m3 V=3,m3 V=2,m3 V=1,m3 Zero Storage Fig. 14: Dual-mode cycle fuel consumption (medium wind case) As one may obtain from Figs , the impact of using even 1,m 3 of storage volume is critical in the reduction of the dual-mode CAES fuel consumption (see for example the 6MW wind power case), by more than 5%, 73% and 92% for the low, medium and high wind potential cases respectively, with the corresponding contribution of the pure CAES cycle obtained by Figs Besides that, as expected, dual-mode fuel consumption becomes zero once hourly load rejections also become zero, since from that point onward, the system relies on the operation of the Wind-CAES scheme only (see also Figs. 6-8). Dual Mode Fuel Consumption (t NG) Dual Mode Fuel Consumption (High Wind Potential Case) V=1,m3 V=9,m3 V=8,m3 V=7,m3 V=6,m3 V=5,m3 V=4,m3 V=3,m3 V=2,m3 V=1,m3 Zero Storage Fig. 15: Dual-mode cycle fuel consumption (high wind case) 5. ECONOMIC EVALUATION OF THE PROPOSED ENERGY SOLUTION Evaluation of the above energy results is undertaken using the economic criterion of long-term electricity production cost, taking into account the input data of Table 2 (for more information on the methodology applied see also [13]). In this context, all three alternative energy solutions are evaluated, i.e. the dual-mode Wind-CAES scheme, the gas-turbine only scheme and finally the wind-farm and gas-turbine parallel operation, with results obtained provided in Figs TABLE 2: MAIN COST PROBLEM INPUTS Wind power cost "c WP " ( /kw) 11 Compressor cost "c c " ( /kw) 4 Gas-turbine cost "c t " ( /kw) 4 Cavern cost "c ss " ( /m 3 ) 25 Wind M&O coefficient "m wind " 2% Gas-turbine M&O coefficient "m gt " 5% Natural gas price "c f "( /tonne) 3 Fuel price escalation rate "e f " 6% Return on investment index "i" 8% M&O cost inflation "g" 4% State subsidy "γ" % Years of operation "n" 2 On top of that, in order to better interpret the economic performance of different dual-mode Wind-CAES configurations, participation of the dual-mode cycle in terms of fuel consumption (in comparison with the total including also fuel consumption of the pure CAES cycle) 6

7 is also given on the right axis of the figures. In this regard, the long-term electricity production cost of the dual-mode Wind-CAES solution presents a gradual increase in the case of the low wind potential as the wind power capacity increases, although in the cases of medium and high wind potential, a minimum optimum point is obtained for N WP =8MW. El. Prod. Cost ( /MWh) Long-Term Electricity Production Cost of Different Energy Autonomous Configurations (Low Wind Potential Case) GT only V=1,m3 V=6,m3 V=1,m3 V=8,m3-DM V=4,m3-DM Wind & GT V=8,m3 V=4,m3 V=1,m3-DM V=6,m3-DM V=1,m3-DM Fig. 16: Economic evaluation results (low wind case) El. Prod. Cost ( /MWh) Long-Term Electricity Production Cost of Different Energy Autonomous Configurations (Medium Wind Potential Case) GT only V=1,m3 V=6,m3 V=1,m3 V=8,m3-DM V=4,m3-DM Wind & GT V=8,m3 V=4,m3 V=1,m3-DM V=6,m3-DM V=1,m3-DM Fig. 17: Economic evaluation results (medium wind case) El. Prod. Cost ( /MWh) Long-Term Electricity Production Cost of Different Energy Autonomous Configurations (High Wind Potential Case) GT only V=1,m3 V=6,m3 V=1,m3 V=8,m3-DM V=4,m3-DM Wind & GT V=8,m3 V=4,m3 V=1,m3-DM V=6,m3-DM V=1,m3-DM Fig. 18: Economic evaluation results (high wind case) DM-CAES Contribution (%) DM-CAES Contribution (%) DM-CAES Contribution (%) At the same time, the gas-turbine-only solution cost is estimated at almost 12 /MWh, which also comprises the more cost-efficient energy solution in case that the wind power capacity exceeds a certain limit. Furthermore, as it derives, the most cost-efficient solution corresponds to the dual-mode Wind-CAES solution with a storage volume of 1,m 3, that however implies extreme levels of the gasturbine cycle participation. On the other hand, as the quality of the local wind potential improves, additional dual-mode Wind-CAES configurations of greater storage capacity become cost-competitive to both the gas-turbineonly and the wind park & gas-turbine solutions, achieving at the same time minimum participation of the dual-mode cycle and thus minimum fuel consumption. To this end, apart from the increase of energy security levels deriving from the reduction of fuel consumption, one should also take into account the fact that solutions determined by low contribution of the gas-turbine cycle are also much less sensitive to the price volatility of natural gas. 6. CONCLUSIONS Based on the development of an energy analysis algorithm for the investigation of dual-mode Wind-CAES configurations, applications results were currently obtained on the basis of different quality wind potential areas examined. To this end, the impact of the local wind potential was reflected on both the energy autonomy levels achieved by the use of a Wind-CAES only scheme and the total fuel consumption required to operate the CAES and gas-turbine cycle. Furthermore, economic evaluation of the alternative solutions investigated designated optimum dual-mode Wind-CAES configurations that however implied maximum contribution of the dual-mode cycle. On the other hand, according to the results, as the local wind potential quality improves, dual-mode Wind-CAES configurations that allow minimum participation of the dual-mode cycle and thus minimum fuel consumption become more costcompetitive, especially in comparison with the gasturbine only solution. As a result, the rational of adopting the proposed solution is illustrated in terms of both longterm electricity production cost and limited fuel consumption, although it must be noted that to obtain a solid conclusion on the economic performance of the proposed solution, a complete sensitivity analysis is also required. 7. REFERENCES (1) Ackermann, T., Andersson, G., Söder, L., Distributed generation: a definition. Electric Power Systems Research 57(3), pp , 21. (2) Kaldellis, J.K., Zafirakis, D., Present situation and future prospects of electricity generation in Aegean Archipelago islands. Energy Policy 35(9), pp , 27. (3) Kaldellis, J.K., Zafirakis, D., Kavadias, K., Technoeconomic comparison of energy storage systems for island autonomous electrical networks. Renewable and Sustainable Energy Reviews 13(2), pp , 29. (4) Denholm, P., Kulcinski, G.L., Life cycle energy requirements and greenhouse gas emissions from 7

8 large scale energy storage systems. Energy Conversion and Management 45(13-14), pp , 24. (5) Denholm, P., Improving the technical, environmental and social performance of wind energy systems using biomass-based energy storage. Renewable Energy 31(9), pp , 26. (6) Bullough, C., Gatzen, C., Jakiel, C., Koller, M., Nowi, A., Zunft, S., Advanced adiabatic compressed air energy storage for the integration of wind energy, Proceedings of the European Wind Energy Conference, London, UK, 24. (7) Dayan, A., Flesh, J., Saltiel, C., Drying of a porous spherical rock for compressed air energy storage. International Journal of Heat Mass Transfer 47(19-2), pp , 24. (8) Greenblatt, J.B., Succar, S., Denkenberger, D.C., Williams, R.H., Socolow, R.H., Baseload wind energy: modeling the competition between gas turbines and compressed air energy storage for supplemental generation. Energy Policy 35(3), pp , 27. (9) Lund, H., Salgi, G., Elmegaard, B., Andersen A.N., Optimal operation strategies of compressed air energy storage (CAES) on electricity spot markets with fluctuating prices. Applied Thermal Engineering 29(5-6), pp , 29. (1) Cavallo, A.J., Controllable and affordable utilityscale electricity from intermittent wind resources and compressed air energy storage (CAES). Energy 32(2), pp , 27. (11) Salgi, G., Lund, H., System behaviour of compressed-air energy-storage in Denmark with a high penetration of renewable energy sources. Applied Energy 85(4), pp , 28. (12) Zafirakis, D., Kaldellis, J.K., Autonomous dual-mode CAES systems for maximum wind energy contribution in remote island networks. Energy Conversion and Management 51(11), pp , 21. (13) Zafirakis, D., Kaldellis, J.K., Economic evaluation of the dual mode CAES solution for increased wind energy contribution in autonomous island networks. Energy Policy 37(5), pp , 29. 8