工控设备plc在氢气产品生产管理上的应用Small stationary reformers for H2 production from hydrocarbons

Small stationary reformers for H 2production from hydrocarbons 5Eugenio Calo`*,Antonella Giannini,Giulia Monteleone ENEA e Italian National Agency for new Technologies,Energy and the Environment Energy Department,C.R.Casaccia,via Anguillarese,30100123Santa Maria di Galeria,Rome,Italya r t i c l e i n f oArticle history:Received 16June 2009Received in revised form 3March 2010Accepted 13March 2010Available online 10April 2010Keywords:Reformer Hydrogen LPGPolymer electrolyte fuel cella b s t r a c tThis paper describes the activities performed in ENEA (Italian National Agency for New Technologies,Energy and Environment)during the last years in order to investigate the hydrogen production from largely available hydrocarbons,such as natural gas [1]and LPG (Liquefied Petroleum Gas),aimed at feeding Polymer Electrolyte Fuel Cells (PEFC)in the 1e 5kW electric power range.Reformers in such a size will be used for small domestic CHP systems (residential,hotels,offices,shopping centres)and for APU on boats and caravans.A significant part of the problems related with residential CHP based on PEFC occurs in the fuel processor [2]:start-up time is to be reduced,efficiency and reliability are to be increased.For this purpose ENEA has tested two LPG fed small reformers with a hydrogen production capacity of about 1Nm 3/h of H 2equivalent,both based on the steam reforming process.The reformer test station was provided with instrumentation in order to evaluate the performances of the reformer in terms of H 2yield,CO content of the syn-gas,thermal and energy balance.The experimental results show the feasibility of such a reformer and indicate the lines for further improvements.In particular the methanation process looks like more suitable than PSA for reformate gas purification in reformers of such a size.The quality of the reformate gas,in term of low CO content,is quite good for feeding PEMFC.The duration and the stability of the start-up time must be improved.The reformer efficiency range is 75e 85%.ª2010Professor T.Nejat Veziroglu.Published by Elsevier Ltd.All rights reserved.1.IntroductionMarket perspectives for small-scale Fuel Processors aimed to feed PEMFC (Proton Exchange Membrane Fuel Cells)forresidential CHP systems,single power generation and small hydrogen filling stations seem to be interesting.With regard to small-scale residential CHP (Combined Heat and Power)the Fuel Processor/PEM Fuel Cell systems have toAbbreviations:APU,auxiliary power unit;CHP,combined heat and power;ENEA,Italian National Agency for new Technologies,Energy and the Environment;FISR,Italian Special Fund for Research;HHV,higher heat value;HIWAR,heat integrated wall reactor;HTS,high temperature shift reaction;LHV,lower heat value;LPG,liquefied petroleum gas;LTS,low temperature shift reaction;MS,mole sieve;PEFC,polymer electrolyte fuel cells;PEM,proton exchange membrane;PEMFC,proton exchange membrane fuel cells;PLC,program-mable logic controller;PNR,National Research Programme;PSA,pressure swinging adsorption;SEM,scanning electron microscope;SOFC,solid oxide fuel cell;WGSR,water gas shift reaction.5Research funded by FISR (Italian Special Fund for Research).*Corresponding author .Tel.:þ390630486231;fax:þ390630483190.E-mail addresses:eugenio.calo@casaccia.enea.it (E.Calo`),antonella.giannini@casaccia.enea.it (A.Giannini),giulia.monteleone@casaccia.enea.it (G.Monteleone).A v a i l a b l e a t w w w.s c i e n c e d i re ct.c o mj o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /h ei n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 35(2010)9828e 98350360-3199/$e see front matter ª2010Professor T.Nejat Veziroglu.Published by Elsevier Ltd.All rights reserved.doi:10.1016/j.ijhydene.2010.03.067face,beside technical development problems(reliability, duration,purity of produced hydrogen,low temperature for useful heat exploitation,etc)even competition of systems based on SOFC(Solid Oxide Fuel Cells)or on internal combustion motor engines.Almost40%of the problems related with small-scale resi-dential PEMFC/CHP systems occur in the fuel processor[2].The most important requirement for stand-alone power generation is independency of the system from electric grid and possibly from any other network such as water or fuels. That means making systems which can be fed by easily available and easily transportable fuels(LPG,gas oil,gasoline, ethanol,liquid fuels in general).For the small hydrogenfilling station the most important requirements,in the long term view,are reliability,duration and cost of produced hydrogen.For a fuel cell-based system,the use of one single hydrogen generation unit must ensure the following features:rapidstart-up,good dynamic response,high-fuel conversion, simple construction and operation,and low cost[3].The choice of a suitable fuel processor and fuel,during the transition phase to a hydrogen economy,are the key aspects to the successful implementation of direct-hydrogen fuel cell systems[4].Fuel processors based on steam reforming appear as the most efficient option for H2rich gas production[5e8].The present paper describes the activities performed in ENEA during the last three years in order to investigate the hydrogen production from a largely available fuel such LPG (Liquefied Petroleum Gas),aimed at feeding Polymer Electro-lyte Fuel Cells(PEFC)in the1e5kW power range.The activities were funded by the national FISR(Italian Special Fund for Research)program.The mainfield of application of the tested devices,once developed,is expected to be that of stand-alone power generation and of local energy networks and apartment buildings.2.Tested reformers and results2.1.First LPG processor(2006e2007)Thefirst LPG Processor to be investigated was provided by Hydrogen/Arcotronics(Italy).The gross size(cm)of the fuel processor,shown in Fig.1,is 80(width)Â40(length)Â40(height).The amount of catalysts used in the different reaction vessels(and therefore the GHSV of reactions)was treated as confidential by manufacturers.LPG is fed to the fuel processor at a pressure of3,5bar (gauge).After a desulfurization unit(150 C e200 C;heated in counter-flow by exhaust gases from the burner,size diameter 34mm,length355mm,sorbent Zinc Cu and Al oxides),the device contains afirst reaction vessel made up by co-axial cylinders enveloped on one another(net size in cm is33of lengthÂ19of external diameter)where,beside the combus-tionflue gas path,there are mainly three different areas of chemical reaction(see Fig.2):1.LPG steam reforming(endothermic)C3H8þ3H2O>3COþ7H2C4H10þ4H2O>4COþ9H2Molar ratio:H2O/C3H85(only for steam reforming)and7 (taking into account reforming and shift),Catalyst:NiO(Johnson Matthey CRG LHR for pre-reforming and Katalco25e4M for reforming),the steam reforming reactor has a useful length of0.20m and an annular section with an equivalent radius of0,0343m, the temperature of reaction is>600 C2.Water Gas Shift Reaction(exothermic)COþH2O>CO2þH2Catalysts:Fe,Cr,Cu oxides(Katalco71e5M)for HTS(300 C e450 C)Cu,Zn oxides(Katalco83-3MX)for LTS(300 C)3.Methanation(exothermic)COþ3H2>CH4þH2OCatalyst:Ni on Calcium aluminate cement(Katalco 11e4MR)the temperature of reaction is300 C,the WGSR and methanation reactor has a useful length of 20cm and an annular section with an equivalent radius of 5,95cm.Outside this reaction vessel,the last part of the processor contains two purification units,based on the mole sieves principle,for alternate adsorption(PSA Pressure Swinging Adsorption)of residual CO.While one of the2mol sieve (MS)units purifies the syn-gas by adsorbing residual CO,the other MS unit is regenerated for180s by means of a LPG stream,and for17s by means of a quota of the reformate gas.Both the streams coming out the regenerating unit are afterwards sent to the burner for combustion.Each of the mole sieves unit is26cm in length and3,5cm in diameter.The pressure drop through each of the molesieves Fig.1e Hydrogen/Arcotronics LPG Processor.i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y35(2010)9828e98359829unit was 0,26bar,and they did not need a compressor,the feeding pressure was 2,5bar gauge,the working temperature 70 C.The design performance (in term of removed CO)was:CO inlet 1.000ppm;CO outlet <50ppm.The design architecture of this LPG Processor was aimed to reduce the overall size of the device and to optimise the heat balance.The heat for the endothermic steam reforming reaction is provided both from the combustion of a small amount of LPG and from the exothermic WGSR (Water Gas Shift reaction)and Methanation reactions.In order to evaluate the performances of the LPG Processor a system for gas analysis is needed.The analysis system is made up by the following instruments:ABB Advance Optima URAS 14(infrared detector)for CO,CO 2,CH 4ABB Advance Optima Caldos 15(thermal conductivity)for H 2ABB Advance Optima electrochemical cell for O 2Varian CP-4900micro GC Molsieve 5A for O 2,N 2,CO,CH 4Varian CP-4900micro GC PoraPlot-U for C 3H 8,CO 2.The experimental performing of this fuel processor showed many key points to be improved:-burner-combustion flue gas tightness -reactants tightness -PSA purificationImproving the burner (Fig.3b)meant mainly to increase the size and the flow rate of the combustion air fan,and to duplicate the LPG inlet in order to control the reformer heating start-up,when a larger amount of LPG is needed.Nonetheless the warming up of the reformer took at least 2e 2,5h (with an average consumption of 2,24kg/h of LPG).Tightening some leakages in the combustion flue gas circuit (Fig.3c)and in the reactants cylinders (Fig.3d)wasn’t enough to ensure the design pressure and the expected performances inside the processor.As the repair would have required a heavy and mostly entirely destructive welding intervention,ENEA decided to check the PSA as a stand-alone device by using a proper gas mixture.2.1.1.Experimental results2.1.1.1.PSA section.PSA size:two cylinders each 26cm in length and 3,5cm in diameter,mole sieves with metal aluminum silicates.Operating temperature 70 C (by means of two electric heaters)P in 3bar gauge,P out 2,7bar gaugeInlet flow 1.000Nl/h (H 2with 1125ppm CO)CO inlet and outlet detected by ABB Caldos and Varian CP-4900micro GC Molsieve 5AExperimental D P:0.15bar (500Nl/h)-0.94bar (2.500Nl/h).The tests performed on the PSA section showed that a time span of 180(plus 17)seconds was enough for nearly complete CO adsorption (CO <10ppm),see Fig.4,while the same time (the two MS unit have to shift from adsorption to purification and vice versa simultaneously)was not sufficient for a complete regeneration of the adsorbent,probably due to an almost equal pressure in adsorption and regeneration phase.The gas-chromatograph method used for the micro GC column Molsieve 5A was: P column:150kPa T ¼80 CCO retention time:1749min.2.1.1.2.Catalyst layers arrangement.Another critical aspectof this processing design was due to the separation of different catalyst layers in the outer cylindrical reaction area (High Temperature WGSR,Low Temperature WGSR,Methanation).An after test SEM (Scanning Electron Microscope)analysis of the catalyst particles that the different catalytic layers had mixed-up with one another.2.2.Second LPG processor (2008)The second LPG Processor,with the commercial name of APS 1000(Fig.5a),was provided by Exergy (former Arco-tronics)(Italy)as a result of the integration of this latter firm and Helbio SA (Greece)into Morphic (Sweden)in the late 2007.The net size (cm)of the fuel processor,shown in Fig.5,is 60(width)Â60(length)Â115(height).Its weight is about 80kg.The hydrogen production is based on a first stage of Steam Reforming,followed by two stages of Water Gas Shift Reaction and two stages of methanation for COpurification.Fig.2e Hydrogen/Arcotronics LPG Processor:chemical reaction area scheme,2a)reactants inlet side 2b)burner inlet side.i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 35(2010)9828e 98359830HELBIO’s solution for the endothermic steam reformer reactor is a heat exchanger type reactor:the Heat Integrated Wall Reactor (HIWAR International Patent PTC 98010080/22.05.98,Fig.6).The reactor consists of one or more tubes inside which reforming takes place on a catalyst deposited in a thin film on the inner surface of each tube.A combustion catalyst is deposited on the outer surface of the tube.In this manner,combustion is controlled very close to where the demand for heat is located while heat transfer is very efficient across the metallic tube wall.This translates to very efficient and compact reactor design [9].LPG is fed at 2bar gauge.The reactions taking place into HELBIO APS 1000Reformer are:-LPG steam reforming (endothermic),800 CC 3H 8þ3H 2O >3CO þ7H 2C 4H 10þ4H 2O >4CO þ9H 2-High Temperature Water Gas Shift Reaction (exothermic),285 CCO þH 2O >CO 2þH2Fig.3e Hydrogen/Arcotronics LPG Processor:3a)removing the burner;3b)improving the burner;3c)sealing the flue gas circuit;3d)leakages from reactantscylinder.Fig.4e Chromatograms (Varian CP-4900micro GC column:Molsieve 5A)of CO content before (1125ppm)and after (10ppm)purification column.i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 35(2010)9828e 98359831-Low Temperature Water Gas Shift Reaction (exothermic),245 CCO þH 2O >CO 2þH 2-First Methanation (exothermic),200 CCO þ3H 2>CH 4þH 2O-Second Methanation (exothermic),165 CCO þ3H 2>CH 4þH 2OInformation about amount of catalysts used in the different reaction vessels,the size and the shape of each reactor (and therefore the GHSV of reactions)was treated as confidential by manufacturers.The Fuel Processor Helbio APS 1000came to ENEA labs in April 2008.During May and June 2008the test rig and the instrumen-tation were set up (Fig.5b and c).In order to calculate mass and energy balances and to evaluate the performances of Helbio APS 1000,gas analysis and mass flow measurements are needed.The analysis and data acquisition system is made up by Analysis systemABB Advance Optima URAS 14(infrared detector)for CO,CO 2,CH 4ABB Advance Optima Caldos 15(thermal conductivity)for H 2ABB Advance Optima electrochemical cell for O 2Data acquisition:Yokogawa Datum Y-XL 100,used for B Reformate gas composition (from ABB Advance Optima) Data acquisition:Yokogawa Datum CX 2000,used for B Reformate gas flow (from Mass Flow Meter Bronkhorst)B Weight of LPG bottle (in continuous,from a scale con-nected to Yokogawa Datum CX 2000)B Weight of demi water bottle (in continuous,from a scale connected to Yokogawa Datum CX 2000)Power-Meter Hameg HM8115-2for auxiliaries absorbed power and power factor measurements.The experimental tests have been conducted in the period June e September 2008.First of all the warming up time has been ascertained.The design start-up time is 1h 200.In Fig.7it can be seen the fast warming of the steam reforming reactor catalytic bed:T reformer and Tref out curves increase rapidly in the first 30min.Afterwards,due to LPG changeable composition and to the back-pressure of the exhaust flue gas circuit,two automatic stops and re-starts occur so that an overall time of 2h 300has been observed for reaching the stabilization of all the reactions (less than 10ppm CO,see Fig.9and a steady H 2content,see Fig.10)of the produced syn-gas.The temperatures were monitored from the reformer PLC mother-board;the main temperatures reported in Fig.7are T reformer :temperature at the reforming catalyst film; T ref out :temperature at the exit of reforming area; T burner out :temperature at the exit of the burner;T burner :temperature of theburner;Fig.5e a)Helbio/Exergy LPG Processor;b)test rig;c)test rig backview.Fig.6e Helbio HIWAR patent.i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 35(2010)9828e 98359832T before1 meth:temperature before thefirst methanator;T after2 meth:temperature after the second methanator.The uneven shape of all the curves,particularly of the T burner and the T reformer curves,is due to the failures and automatic recoveries in the reformer start-up sequence,with subsequent raises and falls of the temperature values.The T after2 meth curve is less influenced by stops and re-startings because this temperature is measured last after all the reactions and shows a bigger response time.2.2.1.Experimental results(“ready to fuel cell”phase)The most meaningful data for calculating the reformer performances are reported hereafter:B Reformateflow rate:1,24Nm3/hB H2concentration in reformate gas:72%B H2flow rate:0,893Nm3/hB Overall LPGflow rate:0,25kg/hB Reactions LPGflow rate:0,138kg/hB Waterflow rate:0,851kg/hFig.7e Temperature paths in Helbio APS1000LPG Processor(from PLCmother-board).Fig.8e CO concentration(ppm)in the outlet reformate gas analyzed by ABB Advance Optima URAS14(collected by Yokogawa Datum Y-XL100).i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y35(2010)9828e98359833BWater/LPG molar ratio:15,8(15if referred to a pure C 3H 8stream)B Temperatures:see Fig.8;the shape of the curves,in particular the T burner and the T reformer curves are due tofailures and automatic recoveries,led by the PLC start-up sequenceB Auxiliaries absorbed power:120W.The syn-gas contains,on a dry base,more than 72%of hydrogen (Fig.9)and about 25%of CO 2and 1,5%of CH 4.The CO content in the reformate gas is less than 1ppm (Fig.8).2.2.2.Fuel processor efficiencyThe Fuel Processor efficiency is calculated as the ratio of useful energy produced (in terms of hydrogen)during a certain time compared to the energy supplied in the same time.At the steady state (“ready to fuel cell”phase)the energy produced as LHV of hydrogen stream is 0.893Nm 3/h Â0,0899kg/Nm 3Â120MJ/kg ¼9,63MJ/h (2,68kW).The energy supplied is made up by:the overall LPG feed:0.25kg/h Â49MJ/kg ¼12,25MJ/h (3,40kW),and the electricity needs of auxiliaries,120W,equivalent in terms of primary energy sources to 261W [10]that is 0,94MJ/h (0,261kW).The Fuel Processor efficiency,based on LHV base,is therefore calculated (all values in MJ/h)as 9,63/(12,25þ0,94)¼73%3.Conclusions and further developmentBased on the experimental results for Fuel Processor an enthalpy balance for a power generator producing 1electric kW has been workedout.Fig.9e H 2concentration in the outlet reformate gas analyzed by ABB Advance Optima Caldos 15(collected by Yokogawa Datum Y-XL100).Fig.10e Helbio/Exergy Fuel Processor/Fuel Cell integrated assembly.i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 35(2010)9828e 98359834Such a power generator will be made up by coupling HEL-BIO APS1000reformer with an EXERGY Fuel Cell with a nominal electric production of1kW.B Reformer efficiency e based on Low Heat Values:at the steady state(“ready to fuel cell”phase)(H2out(kg/h)Â120MJ/kg)/(LPG in(kg/h)Â49MJ/kg)(1,24Â0,72Â0,0899Â120)/(0,25Â49)¼9,63/12,25¼78,6%(*)(*)without accounting the energy for the auxiliaries electric needsThe expected balance for a Fuel Processor-Fuel Cell system has been calculated based on a fuel cell efficiency h fc¼0,4[11]:the expected gross electric power produced by the Fuel Cell is9,63Â0,4¼3,85MJ/h¼1,07kWTo obtain the net electric power we have to subtract the auxiliaries needs for the reformer section(120W)and for the Fuel Cell and DC/AC inverter section(100W);the net electric power is therefore:1,07À(0,120þ0,100)¼0.85kW(3,06MJ/h)and the overall expected electric efficiency,attainable for such an equipment,is(values in MJ/h):3,06/12,25¼25%(based on Low Heat Value of LPG and taking into account all the auxiliary electric needs).This value is consistent with the design value for both in network and stand-alone performance.As a further development the integrated Fuel Processor Fuel Cell system(Fig.10)will be tested in ENEA laboratories during2009.The outlines for improvements aimed at industrialization of the system are:-Reducing times for(cold)start-up-Reducing size and weight of some components-Reducing the need of electric power consumption for auxiliaries(reducing instrumentation and controls)-Improving heat exchanges and H2recovery from anodic flue gas-Recovering(by re-circulation)the anodic exhaust(con-taining about20%of produced H2,plus some CH4)from PEM stack to feed the burner-Reducing LPG inlet to burner e after the start-up phase:a simulation modeling showed it could be lowered of10%without any problem.AcknowledgementsWe are grateful to:-EXERGY Fuel Cells /-HELBIO SA http://www.morphic.se/en/Helbio/This work has been realised with the contribution of FISR (Italian Special Fund for Research).The fund is aimed at financing strategic research themes,as stated within Italian National Research Programme(PNR).r e f e r e n c e s[1]Calo`E,Cruciani C,Galli S,Monteleone G.Small stationaryreformer for hydrogen production progress activities in ENEA Hypothesis V;7e10September2003[2]Aki Hirohisa.The penetration of micro CHP in residentialdwellings in Japan.IEEE;2006.[3]Ersoz A.Investigation of hydrocarbon reforming processesfor micro-cogeneration systems.International Journal ofHydrogen Energy;October2008.[4]Larminie J,Dicks A.Fuel cell systems explained.2nd ed.;2003.[5]Ogden JM.Review of small stationary reformers for hydrogenproduction.Task Development Workshop;2001.[6]Recupero V,Pino L,Vita A,Cipitı`F,Cordaro M,Lagana`M.Development of a LPG fuel processor for PEFC systems:laboratory scale evaluation of autothermal reforming andpreferential oxidation subunits.International Journal ofHydrogen Energy2005;30:963e71.[7]Laosiripojana N,Assabumrungrat S.Hydrogen productionfrom steam and autothermal reforming of LPG over highsurface area ceria.Journal of Power Sources2006;158:1348e57.[8]Cipitı`F,Pino L,Vita A,Lagana`M,Recupero V.Performance ofa5kWe fuel processor for polymer electrolyte fuel cells.International Journal of Hydrogen Energy2008;33:3197e203.[9]http://www.morphic.se/en/Helbio/Technology/Heat-Integrated-Wall-Reactor-HIWAR/.[10]Italian authority for electric energy and gas.AEEG;2008.data.[11]Cicconardi SP,Granati M.ENEA-MSE(Ministry forEconomical development)“Steady-state and dynamicsimulation for a5kW CHP system”.Italy:University ofCassino;March2009.RS/2009/179.i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y35(2010)9828e98359835。