Document Sample

Proc. of the 5th WSEAS/IASME Int. Conf. on Electric Power Systems, High Voltages, Electric Machines, Tenerife, Spain, December 16-18, 2005 (pp520-525) Unitised Regenerative Fuel Cells for Stand-Alone Photovoltaic Generation Systems D. ARDITO, S. CONTI, S. RAITI, U. VAGLIASINDI Dipartimento di Ingegneria Elettrica, Elettronica e dei Sistemi Università degli Studi di Catania Viale A. Doria, 6, 95125 Catania ITALY Abstract – Photovoltaic (PV) stand-alone applications are able to provide electricity to isolated loads in remote areas and they are installed particularly where grid extensions would be uneconomical. However, the limitation in power availability of PV systems due to the variability of solar radiation requires the use of storage systems in order to supply loads with adequate reliability levels. The storage systems have to store a great amount of energy to be maintained for quite long time periods with small losses. This is quite difficult to be achieved by using electrochemical batteries and the use of hydrogen in regenerative fuel cells as a means for energy storage can represent a solution to reach the aforesaid goals. This paper deals with the use of Unitised Regenerative Fuel Cells (URFC) in the realization of stand-alone PV generation systems. The study of the generation system with solar hydrogen storage will be carried out using analytical models to represent the efficiency of each component in order to assess the capability of the generation system to supply its load with an adequate reliability level in terms of Loss Of Load Probability (LOLP). In this perspective, a comparison between different storage technologies such as Regenerative Fuel Cells (RFC) and Unitised Regenerative Fuel Cells will be presented. The performance of a storage system based on electrochemical batteries will be taken as reference. Key-Words – Renewable Energy, Photovoltaics, Hydrogen Storage System, Unitised Regenerative Fuel Cells. 1 Introduction Research in this field proceeds in the development of The installation of stand-alone renewable energy new technologies to produce hydrogen from water generation systems, such as photovoltaic (PV), at electrolysis. These technologies are, e.g., the those sites were meteorological conditions are Unitised Regenerative Fuel Cells (URFC) [2], [3]. favourable, can bring great benefits in terms of both Usually Fuel Cells (FC) are employed for energy costs and reliability. In fact, stand-alone applications generation in Distributed Generation (DG) due to are able to provide electricity to isolated loads in their high efficiency, reliability and environmental remote areas and they are installed particularly compatibility. Further, FC can play the role of where grid extensions would be uneconomical. Then, energy storage systems. To accomplish this FC need generally speaking, climatic conditions and grid to be coupled with an electrolyser (EZ), which is a supply availability have a basic influence on the hydrogen generator device, to realize the economic evaluation of renewable energy Regenerative Fuel Cell (RFC) system. In practice, installations with respect to other solutions. the RFC system uses two separate cell stacks: an However, the limitation in power availability of electrolysis cell stack (EZ) to produce hydrogen and photovoltaic systems due to the variability of solar a separate FC stack to generate electric power from radiation calls for the use of storage systems in order stored hydrogen. to supply loads with adequate reliability levels. The The URFC refines this concept by using the cell storage systems have to store a great amount of electrodes to perform both the EZ function and the energy to be maintained for quite long time periods FC function. Hence, the URFC system uses a single with small losses. This is quite difficult to be reversible cell stack to alternately produce hydrogen achieved by using electrochemical batteries due to from electrical energy and regenerate electrical low efficiency and self-discharge. At present, the use energy on demand from stored hydrogen. The URFC of hydrogen intended as a means for chemical systems have lighter weight and smaller physical storage and transfer of solar energy seems to be a size than those systems that employ separate cell solution to overcome the aforesaid limitations [1]. stack [4]. Proc. of the 5th WSEAS/IASME Int. Conf. on Electric Power Systems, High Voltages, Electric Machines, Tenerife, Spain, December 16-18, 2005 (pp520-525) In this paper, the operation of a PV stand-alone PV LOAD generation system with solar hydrogen storage will be investigated by using analytical models to represent the efficiency of each component in order to assess the capability of the generation system of STORAGE supplying its load with an adequate reliability level in terms of Loss Of Load Probability (LOLP). Fig. 1. Simplified block scheme of a PV stand-alone Further, a comparison between different storage generation plant with storage systems technologies such as Regenerative Fuel Cells (RFC) and Unitised Regenerative Fuel Cells (URFC) will RFC system be carried out by taking as reference the performance H2O of a storage system based on electrochemical batteries. H2 in PRFC EZ Tank H2 out FC PRFC H2 2 Schemes of a stand-alone PV O2 /air system with storage a) The simplified block diagram of a PV stand-alone URFC system generation plant with storage system is shown in Fig. H2O 1. H2 in The storage system based on FC technology is PURFC shown in Figg.2 a) and b), respectively, for RFC and Tank H 2 EZ/FC out PURFC URFC. Obviously, the system components of Fig. 1 H2 can not be connected directly to each other for the O2 /air following reasons: • different voltage levels in the system; FC operation • control and possible optimisation of global EZ operation efficiency would be impossible; b) Fig. 2. Schemes for RFC and URFC systems • necessity to convert cc waveforms into ac ones. As a consequence, it is necessary to employ DC/DC and DC/AC converters, with different control schemes according to the various storage PV MPPT DC/AC LOAD technologies, in order to provide power conditioning, efficiency optimisation and subsystems coupling [5], [6]. This work will deal with the plant typologies shown in Fig. 3, where: URFC TANK - MPPT = Maximum Power Point Tracker; - DC/AC = inverter; a) - DC/DC = converter; - GC = Gas Compressor. PV MPPT DC/AC LOAD EZ FC 3 Energy flows assessment The aim of this Section is to analyse the energy GC TANK flows within the solar hydrogen system. To do this, analytical models for the various system components b) in the considered configurations (Fig. 3) have been developed. The basic scheme used for the energy PV MPPT DC/AC LOAD flows assessment is shown in Fig. 4. We define the following quantities: DC/DC DC/DC • λi the irradiance on a surface with a given BATTERY inclination to the horizontal plane during the i-th hour (i=1…24) [kW/m2]; c) • PLi = load power demand during i-th hour Fig. 3. Configurations with by different storage (i=1…24) [kW]; systems: a) URFC, b) RFC, c) Electrochemical batteries • A is the array surface area [m2]. Proc. of the 5th WSEAS/IASME Int. Conf. on Electric Power Systems, High Voltages, Electric Machines, Tenerife, Spain, December 16-18, 2005 (pp520-525) 3.1 Load energy demand Obviously, a portion of this surplus energy can be Since load power demand PL is considered constant stored (ESj). This is due to the storage system during the i-th hour, it can be assumed that PLi=ELi efficiency, so that: (where EL is the load energy demand during the E = ηc E (6) considered hour of the year, i = 1,…, 8760). Sj Sj SURj The yearly load energy demand, ELy, is given by: where η Sj is the storage system efficiency in charge c 8760 E Ly = ∑ E Li (1) mode, variable with ESURj . i =1 The expression of η Sj depends on the technology c 3.2 Photovoltaic energy used to realise the storage system: • URFC → η Sj = ηURFCj c EZ The energy produced (superscript p) by the PV array during the i-th hour is: • RFC → η Sj = η RFCj ⋅ ηcomp c EZ p E PVi=A ⋅ ηPVi ⋅ λi (2) • Battery → η Sj = η BATTj ⋅ η DC / DC c c where: ηPVi is the efficiency of the PV array, variable with Discharge mode: definition of “deficit energy” and the hour and solar irradiance λi . “provided energy”. The hourly PV energy actually available At node N of Fig. 4, during the k-th hour, the (superscript a) downstream from the MPPT is given following inequality holds: by: a E PVk < E Lk /ηinv a p (7) E PVi =ηMPPT ⋅ E PVi (3) This means that the PV energy made available is where ηMPPT is the efficiency of the MPPT. lower than load energy demand. We have a deficit energy, EDEFk , given by: N a E DEFk = E Lk / ηinv - E PVk = PV MPPT DC/AC LOAD (8) = E Lk / ηinv - Aη MPPT η PVk λk STORAGE Obviously, the energy actually provide to the storage system, EPROVk , will be greater than the Fig. 4. Basic scheme used for energy flows assessment deficit energy. This is due to the storage system efficiency, so that: 3.3 Storage system energy E Two operation modes for the storage system have EPROVk = DEFCk (9) been identified: the charge mode (superscript c) and d ηSk the discharge mode (superscript d). where η d Sk is the storage system efficiency in Accordingly, the hours of the year will be discharge mode, variable with EDEFk . distinguished in “j” hours (charge hours), when PV production exceeds load demand, and “k” hours Similarly to the previous case, the expression of (discharge hours), when PV production is lower than ηSk depends on the technology used to realise the d load demand. The two operation modes are storage system: • URFC → ηURFCk characterised by different hourly efficiencies, FC respectively, η Sj and ηSk . c d • RFC → ηRFCk FC • Battery → η BATTk ⋅ ηDC / DC d Charge mode: definition of “surplus energy” and “stored energy”. At node N of Fig. 4, during the j-th hour the following inequality holds: 4 Efficiency analytical models of EPVj > ELj / ηinv a (4) system components where ηinv is the DC/AC inverter efficiency. The efficiency analytical models for each component This means that the PV energy made available have been derived from experimental data. In the exceeds the load energy demand. We have a surplus models the efficiency is expressed as a function of energy, ESURj , given by: the component input / output power: ESURj = EPVj - E Lj / ηinv = Aη MPPTη PVj λ j - E Lj / ηinv a (5) η=η(P) (10) Proc. of the 5th WSEAS/IASME Int. Conf. on Electric Power Systems, High Voltages, Electric Machines, Tenerife, Spain, December 16-18, 2005 (pp520-525) As previously highlighted, the hourly input or η RFC = 0.87289 − 1.38619 ⋅ ( pRFC ) + EZ in output power is assumed constant so that, for each in in (19) given hour energy is numerically equal to power: + 2.55711 ⋅ ( pRFC ) 2 − 1.41009 ⋅ ( pRFC )3 P=E. Consequently, expression (10) is equivalent to η RFC = 0.70721 − 0.74683 ⋅ ( pRFC ) + FC out the following: out out (20) η=η(E) (11) + 1.13758 ⋅ ( pRFC ) 2 − 0.77097 ⋅ ( pRFC )3 The coefficients have been derived from 4.1 PV efficiency model experimental data provided by CNR - ITAE The PV hourly efficiency is expressed as a function (National Research Council - Institute for advanced of global solar irradiance λi : energy technologies), in Messina (Italy). 7.9 ⋅ 10 −4 η PVi = 9.55 ⋅ 10 − 2 − (12) 4.4 Battery efficiency model λi The efficiency model has been developed on the The coefficients of the expression have been ground of data provided by [9] and [10]. obtained from experimental measures [7]. As said before, it is necessary to take into account a charge efficiency ( η BATT ) and a discharge c 4.2 URFC efficiency model efficiency ( η BATT ), d respectively, functions of The efficiencies in EZ operation mode ( η EZ URFCj ) and p in and p out , that are the relative values of input BATT BATT in FC mode ( ηURFCk ) are, respectively, expressed as FC in out and output power PBATT and PBATT , referred to peak in out functions of p URFC and p URFC , that are the relative EZ FC powers PBATTp and PBATTp , i.e.: values of input and output power P in and P out , in in EZ URFC URFC pBATT = PBATT / PBATTp (21) EZ FC referred to peak powers P URFCp and P URFCp , i.e.: out out FC pBATT = PBATT / PBATTp (22) in in EZ pURFC =P URFC /P URFCp (13) The expressions for battery efficiency obtained out out FC pURFC =P URFC /P URFCp (14) are the following: ( η BATT = ⎡1- 0.2 pBATT ⎤ (1- 0.1)) c in 2 ( N -1) The obtained efficiency expressions are the (23) ⎢ ⎣ ⎥ ⎦ following: η BATT = ⎡1- 0.6 ( pBATT ) ⎤ (1- 0.1) 2 ( N -1) ηURFC = 0.89265 − 1.09878 ⋅ pURFC + EZ in d out (24) (15) ⎢ ⎣ ⎥ ⎦ in in + 1.81734 ⋅ ( pURFC ) 2 − 1.04471 ⋅ ( pURFC )3 where N=1…10 is the battery year of life that must ηURFC = 0.83465 − 0.69088 ⋅ pURFC + FC out be taken into account in order to consider the out out (16) reduction in efficiency due to self-discharge and + 0.89368 ⋅ ( pURFC )2 − 0.64195 ⋅ ( pURFC )3 electrodes degradation. This is important because the The coefficients have been derived from battery life-cycle is much shorter than that of FC. experimental measures carried out on a test URFC of type #9804A (produced by Proton Energy Systems Inc., Connecticut, USA, and tested by LLNL - 5 Loss of Load Probability Lawrence Livermore National Laboratory - The efficiency analytical models presented in the California, USA) [8]. previous Sections for each component in the various configurations will be employed to assess the 4.3 RFC efficiency model capability of the generation system to supply the load The EZ efficiency ( η RFCj ) and the FC efficiency EZ with an adequate reliability level. To do this we will define the known reliability ( η RFCk ) are, respectively, expressed as functions of FC index called LOLP (Loss of Load Probability) as: in out pRFC and pRFC , that are the relative values of input in out LOLP = ∑ hLL and output power PRFC and PRFC , referred to peak H (25) EZ FC powers PRFCp and PRFCp , i.e.: where the nominator is the sum of the overall “loss in in EZ pRFC = PRFC / PRFCp (17) of load hours” (indicated by hLL) - discharge hours - during which the storage system is not able to meet out out FC pRFC = PRFC / PRFCp (18) the load demand; the denominator, H, is the total The obtained efficiency expressions are the number of hours in the year (8760). following: Proc. of the 5th WSEAS/IASME Int. Conf. on Electric Power Systems, High Voltages, Electric Machines, Tenerife, Spain, December 16-18, 2005 (pp520-525) 6 Monte Carlo Simulation In physical terms, the quantity he represents the LOLP calculation for the various system number of hours during which the storage system is configurations has been carried out by means of a able to meet a load demand equal to PLav. software tool developed by the Authors on The results obtained are referred to an ideal load MATLAB® platform, based on Monte Carlo (MC) diagram (shown in Fig. 5) which brings about the method. This method allowed to obtain a realistic maximum energy storage. assessment of system reliability by using the PLmax statistical variations of load demand and PV production. This has been done by means of PLav appropriate statistical models for power demanded by the load and produced by the PV generator [11]. PL (kW) Once the statistical models are defined in terms of 0.4PLmax probability density functions (pdfs) the procedure involves repeating the simulation using each time 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 (hour by hour) a particular value of the random hours (h) variables (load demand and PV production), generated according to the corresponding pdfs. Fig. 5. Ideal daily load diagram For each hourly simulation it is possible to assess whether the considered hour is a “loss of load hour” This is because the concern is on the operating or not. The simulations are then extended to the condition in which the storage system assumes the overall year, thus obtaining the sum of the loss of most critical role from the reliability viewpoint. This load hours and then the value of the LOLP for that condition is when the generation diagram has its year (LOLPy). maximum during minimum load hours. Hence the To ensure a reasonable accuracy of the surplus energy is maximum with respect to load calculation performed by the Monte Carlo method, demand. an appropriate number of years (Y) is to be Graphs of Figg. 6, 7, 8 and 9 show that the PV considered. The final result will be the average generation system with hydrogen-based storage LOLP value: technology is more reliable than the system with Y electrochemical batteries. Further, the configuration ∑ LOLP y =1 y with URFC has a lower LOLP (reduced by a 10%) LOLP = (26) than the configuration with RFC. Consequently, the Y URFC, besides being advantageous in terms of light weight and small physical size, ensures higher 7 Numerical results reliability levels than those systems that employ This section presents the results of the analysis separate cell stacks (RFC). carried out to assess the reliability in terms of LOLP index for the three configurations characterized by 1 LOLP LOLPurfcURFC different storage systems. In particular the values of LOLP LOLPrfc RFC 0,9 LOLPURFC, LOLPRFC LOLPBATT will be reported in LOLPbattBATT LOLP LOLP the graphs of Figg. 6, 7, 8 and 9 as a function of p 0,8 (adimensional) which is defined as the relative value 0,7 of PPvpeak (kW) referred to the daily average value of 0,6 load demand, PLav (kW): 0,5 PPV peak 0 2 4 6 8 10 12 14 16 18 20 22 p= (27) p PL av Fig. 6. LOLP graphics with he=1 where PPvpeak is the value of power generated by the PV system in standard conditions (Solar 1 LOLPURFC LOLPurfc Irradiation=1kW/m2 and Cell Temperature=25ºC). 0,9 LOLPRFC LOLPrfc The aforesaid graphs are characterised by 0,8 LOLPBATT LOLPbatt different capacities, Es (kWh), of the storage system 0,7 LOLP 0,6 expressed in terms of “equivalent hours” (he), 0,5 defined as: 0,4 ES 0,3 he = (28) 0 2 4 6 8 10 12 14 16 PL av p Fig. 7. LOLP graphics with he=12 Proc. of the 5th WSEAS/IASME Int. Conf. on Electric Power Systems, High Voltages, Electric Machines, Tenerife, Spain, December 16-18, 2005 (pp520-525) 1 LOLP LOLPurfcURFC References: 0,9 LOLP LOLPrfc RFC [1] Mitlitsky F., Myers B. and Weisberg A.H. 0,8 LOLP LOLPbattBATT (LLNL, California), “Regenerative Fuel Cell 0,7 LOLP 0,6 System R&D”, Proceedings of the 1998 U.S. 0,5 DOE Hydrogen Program Review (NREL/CP- 0,4 570-25315). 0,3 0,2 [2] F. Mitlitsky, B. Myers and N.J Colella. 0 2 4 6 8 10 p 12 14 16 18 20 22 “Unitized regenerative fuel cell for solar rechargeable aircraft and zero emission Fig. 8. LOLP graphics with he=24 vehicles”, Fuel Cell Seminar, San Diego Nov/Dic 1994. 1 LOLP LOLPurfcURFC [3] J.F. McElroy, “Unitized Regenerative Fuel Cell 0,9 0,8 LOLP LOLPrfc RFC Energy Storage Systems For Aircraft And 0,7 LOLP LOLPbattBATT Orbital Applications” UTC Hamilton Standard 0,6 Div., Rept BD94-02, Mar 1994.K. LOLP 0,5 0,4 0,3 [4] A. Burke, “High Energy Density Regenerative 0,2 Fuel Cell Systems for Terrestrial Applications”, IEEE AES Systems Magazine, December 1999. 0,1 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 [5] H. Solmecke, O. Just, D. Hackstein, p “Comparison of Solar Hydrogen Storage Fig. 9. LOLP graphics with he=48 Systems with and without Power-Electronic DC-DC Converters”, Renewable Energy (Pergamon), 19 (2000), pp. 333-338. [6] J. Appelbaum, “The Operation of Loads 4 Conclusion Powered by Separate Sources or by Common In this paper, the operation of a PV stand-alone Source of Solar Cell”, IEEE Transactions on generation system with solar hydrogen storage has Energy Conversion, Vol. 4, No. 3, September been investigated by using analytical models to 1989. represent the efficiency of each component in order [7] M. Guerra, A. Sarno, S. Raiti, G. Tina, to assess the capability of the generation system to “ENEA's facilities to test small-size inverters supply the load with an adequate reliability level. for low voltage grid-connected PV systems”, This allowed a comparison between different 16th European Photovoltaic Solar Energy storage technologies such as Regenerative Fuel Cells Conference and Exhibition, EPSE'2000, 1-5 (RFC) and Unitised Regenerative Fuel Cells May, 2000, Glasgow, Scotland (URFC), carried out by taking as reference the [8] F. Mitlitsky, A.H. Weisberg, B. Myers, performance of a storage system based on “Vehicular Hydrogen storage using Lightweight electrochemical batteries. Tanks (Regenerative Fuel Cell systems)”, Proc. The aforesaid comparison resulted in the higher of the 1999 U.S. DOE Hydrogen Program reliability level of the PV stand-alone generation Review (NREL/CP-570-26938). system with hydrogen storage as referred to the use [9] H. Senoh, Y. Hara, H. Inoue, C. Iwakura, of electrochemical batteries. Further, as for the “Charge Efficiency of Misch Metal-Based hydrogen storage system, the URFC guarantees Hydrogen Storage Alloy Electrodes at higher reliability level than those systems that Relatively Low Temperatures”, Electrochimica employ separate cell stacks (RFC). Acta (Pergamon), 46 (2001), pp. 967-971. This makes the URFC technology very attractive [10] M. Pedram, Q. Wu, “Design Considerations for to store energy in the form of hydrogen and to Battery-Powered Electronics”, Proceedings of produce electrical energy from the stored hydrogen the 36th ACM/IEEE Conference on Design as well. Automation (ACM Press, June 1999). However, URFC technology still involves [11] S. Conti, T. Crimi, S. Raiti, G. Tina, U. uncompetitive costs as compared to RFC technology Vagliasindi, “Probabilistic Approach to Assess (URFCs are only employed in space applications the Performance of Grid-Connected PV where the high costs are determined also by the use Systems”, Proceedings of 7th International of precious materials). Conference on Probabilistic Methods applied to Power Systems, September 22-26., Naples, Italy.

DOCUMENT INFO

Shared By:

Categories:

Tags:
solar energy, fuel cells, wind power, pem fuel cell, united kingdom, crop production, power systems, solar panels, renewable energy sources, change legislation, home energy, hydropower project, press release, green energy, alternative energy

Stats:

views: | 25 |

posted: | 6/12/2010 |

language: | English |

pages: | 6 |

OTHER DOCS BY xfo14057

How are you planning on using Docstoc?
BUSINESS
PERSONAL

By registering with docstoc.com you agree to our
privacy policy and
terms of service, and to receive content and offer notifications.

Docstoc is the premier online destination to start and grow small businesses. It hosts the best quality and widest selection of professional documents (over 20 million) and resources including expert videos, articles and productivity tools to make every small business better.

Search or Browse for any specific document or resource you need for your business. Or explore our curated resources for Starting a Business, Growing a Business or for Professional Development.

Feel free to Contact Us with any questions you might have.