Porous burners for lean-burn applications Susie Wood , Andrew T. Harris Laboratory for Sustainable Technology, School of Chemical and Biomolecular Engineering, The University of Sydney, NSW 2006, Australia article info Article history: Received 27 November 2006 Accepted 21 April 2008 Available online 9 June 2008 Keywords: Porous burners Low-grade fuels Lean-burn applications Ultra-lean methane combustion Porous materials Burner design Burner performance abstract We review research on lean methane combustion in porous burners, with an emphasis on practical aspects of burner design and operation and the application of the technology to real-world problems. In particular we focus on ‘ultra-lean’ combustion, where the methane concentration is actually at or below the lean flammability limit for a free flame (5% methane by volume in air). Porous burners are an advanced combustion technology whereby a premixed fuel/air mixture burns within the cavities of a solid porous matrix. They are capable of burning low-calorific value fuels and very lean fuel/air mixtures that would not normally be flammable, potentially allowing the exploitation of what would otherwise be wasted energy resources. Possible lean-burn applications include the reburn of exhaust gases from existing combustion systems, and the mitigation of fugitive methane emissions. Porous burners operate on the principle that the solid porous matrix serves as a means of recirculating heat from the hot combustion products to the incoming reactants. This results in burning velocities higher than those for a free flame, as well as extended lean flammability limits. Burner performance is also characterised by low emissions of combustion related pollutants and stable operation over a wide range of fuel concentrations and flow rates. Stable combustion of methane/air mixtures below the conventional lean limit has been observed by a number of researchers; in one study the combustion of a mixture with a fuel concentration of only 1% was reported. A number of design considerations are important as regards optimising burner performance for lean-burn applications. Foremost among these is the selection of a suitable material for the porous matrix. Possibilities include packed beds of alumina spheres or saddles, and reticulated foams made of silicon carbide or high temperature metal alloys. Other potentially significant design issues include the length of the porous bed, the use of ‘multi-section’ designs where different porous materials are used in each section, the incorporation of external heat exchangers to supplement the heat recirculation provided by the porous matrix, and the ability to operate the burner at elevated pressures. There is an extensive body of research relating to porous burners, comprising experimental and numerical investigations. However the majority of previous studies have been directed towards the use of porous burners for radiant heating applications rather than for the combustion of low-calorific value fuels. Consequently there is a lack of reliable data relating specifically to ultra-lean combustion. We identify specific areas where further research is required to progress this field. These include the influence on burner performance of the design considerations listed above, the stability of the combustion process to fluctuations in fuel concentration and flow rate, the development of reliable models specifically for ultra-lean combustion in practical burners, and the investigation of issues relating to scale-up and commercial application. & 2008 Elsevier Ltd. All rights reserved. Contents 1. Introduction ...................................................................................................... 668 2. Burner operating principles .......................................................................................... 668 2.1. Excess enthalpy combustion ................................................................................... 668 2.2. Combustion in a porous medium ............................................................................... 669 2.2.1. Heat recirculation .................................................................................... 669 2.2.2. Flame stabilisation .................................................................................... 670 ARTICLE IN PRESS Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/pecs Progress in Energy and Combustion Science 0360-1285/$ - see front matter & 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.pecs.2008.04.003 Corresponding author. Tel.: +612 9036 6244; fax: +612 93512854. E-mail address: [email protected] (S. Wood). Progress in Energy and Combustion Science 34 (2008) 667– 684
Transcript
ARTICLE IN PRESS
Progress in Energy and Combustion Science 34 (2008) 667– 684
Contents lists available at ScienceDirect
Progress in Energy and Combustion Science
0360-12
doi:10.1
� Corr
E-m
journal homepage: www.elsevier.com/locate/pecs
Porous burners for lean-burn applications
Susie Wood �, Andrew T. Harris
Laboratory for Sustainable Technology, School of Chemical and Biomolecular Engineering, The University of Sydney, NSW 2006, Australia
a r t i c l e i n f o
Article history:
Received 27 November 2006
Accepted 21 April 2008Available online 9 June 2008
Current concerns over climate change, energy security andrising fossil fuel prices have led to a growing interest in findingalternative and renewable energy sources. Processes involving therecovery of what would otherwise be wasted energy from low-grade fuels might therefore be anticipated to attract increasedattention. Such processes include energy recovery from landfillgas, combustion of biomass derived syngas, reburn of exhaust gasfrom existing combustion systems, and mitigation of fugitivemethane emissions such as coal mine ventilation air. Theadvantages of burning low-calorific value fuels can include anincrease in the energy efficiency of existing energy generationsystems, a decrease in the consumption of conventional fossilfuels, and a reduction in greenhouse gas emissions.
Porous burners are one possible technology for the combustionof very lean fuels that might not otherwise be flammable. They arean advanced combustion technology whereby a premixed fuel/airmixture burns within the cavities of a solid porous matrix, ratherthan as a free flame at the burner exit, as is the case withconventional gas burners. The porous matrix serves as a means ofrecirculating heat from the hot combustion products to theincoming reactants, leading to higher flame speeds and extendedflammability limits.
Porous burners are already commercially available [1,2], andfind application in fields including space and water heating, metalheat-treating, coating and paint drying, glass and chemicalprocessing, paper and wood drying, and food processing [3–5].Furthermore, since as early as the end of the 19th century [6], butparticularly over the last three decades, combustion in porousmedia has been the subject of a significant amount of researchand development, with a number of useful reviews published onthe topic [3,6–14].
However, the combustion of very low-calorific value fuelsusing porous burners is a field that remains relatively unexplored(though in a recent development in this area, the possibility ofusing porous burners to combust landfill gas and low-calorificsyngas from waste pyrolysis was investigated [15]), and theuse of the technology specifically for such lean-burn applica-tions has yet to be demonstrated beyond the laboratory- andpilot-scale.
This review will examine lean methane combustion in porousburners. Aspects of burner design and performance of practicalsignificance to lean-burn applications, rather than the funda-mental mechanisms of the combustion process, will be empha-sised. We are particularly interested in what may be described as
‘ultra-lean’ combustion, where the methane concentration isactually at or below the lean flammability limit for a free flame(5% methane by volume in air [16]).
The use of alternative fuels such as LPG, hydrogen, hydrogensulphide and syngas, liquid fuels such as heptane, kerosene,methanol and petrol, or suspended solid particles such as coaldust, will not be covered, although these have all been thesubjects of recent research [16–30]. Neither will oxy-fuelcombustion [31] be considered, nor the use of porous mediareactors for processes such as syngas or hydrogen production[27,32–37] or NOX reburning [38]. Additionally, the scope of thereview will be limited to homogeneous combustion in inertporous media—research relating to catalytic combustion, orindeed systems involving the combustion of the porous bed itself,will not be included.
The structure of this review will be as follows: First, the basicprinciples governing combustion in a porous medium will beexplained, and the common characteristics of porous burnerssummarised. Second, specific examples of burner performancepreviously reported in the literature will be provided. Third, issuesrelating to porous material selection and other aspects of burnerdesign will be explored. Finally, the current status of research inthe field will be reviewed, and some potential topics for futurework identified.
2. Burner operating principles
2.1. Excess enthalpy combustion
For a laminar premixed flame the adiabatic flame temperaturecan be defined as the theoretical temperature obtained if all theheat released by the reaction is used to raise the temperature ofthe combustion products [39]. It is therefore solely dependent onthe initial reactant composition, in other words on the heatingvalue of the fuel and the fuel/air ratio. However, if a means can befound of recirculating some of the heat from the hot combustionproducts to the cold reactants, whilst at the same time avoidingdilution of the reactants with the products, it should theoreticallybe possible to obtain flame temperatures in excess of the adiabaticflame temperature of the initial fuel/air mixture [40].
The term ‘excess enthalpy’ burning is used to describe thisprocess of ‘borrowing’ enthalpy from the combustion products topreheat the incoming reactants [41]. Because it is characterised byflame temperatures and burning velocities greater than thecorresponding adiabatic flame temperature and laminar burning
ARTICLE IN PRESS
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684 669
velocity it is also commonly referred to as ‘super-adiabatic’combustion.
The flammability limits for a fuel/air mixture are defined as thefuel concentrations within which a self-sustaining flame can form[39]. Because the lean flammability limit of a fuel/air mixturedecreases as the initial temperature of the mixture increasesaccording to the Burgess-Wheeler law [42], excess enthalpyburning can lead to a reduction in this lower limit. Hardesty andWeinberg [41] describe an idealised thermodynamic model ofexcess enthalpy burning and show that in moving towards thecombustion of increasingly lean mixtures, heat losses from thesystem will become controlling, ultimately determining the leanlimit that can be achieved.
Weinberg and colleagues suggested a number of schemes forthe practical realisation of excess enthalpy burning in heatrecirculating combustors involving various configurations of heatexchanger tubing surrounding the combustion chamber. Two ofthese are illustrated in Fig. 1. In experiments using the ‘doublespiral’ burner, they reported stable combustion for methaneconcentrations as low as 1.6% [41,43,44].
2.2. Combustion in a porous medium
An alternative means of achieving excess enthalpy combustionis to insert a porous solid with superior heat transfer propertiesinto the combustion chamber. This provides a means of recircu-lating the heat internally. Rather than having an external heatexchanger surrounding the combustion chamber, the combustiontakes place within the heat exchanger itself. This idea forms thebasis of porous burner operation.
The concept was first demonstrated analytically by Takenoand Sato [45]. Subsequent investigations by Takeno and collea-gues at the University of Tokyo confirmed their predictions
Fig. 1. Examples of heat recirculating burners [43].
Fig. 2. Heat transfer processe
experimentally [46,47]. Using a burner combining both a com-bustion chamber containing a porous medium with an arrange-ment of heat exchanger tubes surrounding it, they were able tosustain combustion at flow velocities higher than could beattributed to the burning velocity of the externally preheatedgases alone—the difference was credited to the contribution ofthe internal heat recirculation provided by the porous solid.
2.2.1. Heat recirculation
Heat recirculation in a porous medium involves a combinationof all three modes of heat transfer—conduction, convection andradiation, as identified in Figs. 2 and 3a. The process can besummarised as follows: Downstream of the reaction zone, the gasis hotter than the solid, and so heat is transferred convectivelyfrom the hot combustion products to the porous matrix; the hotsolid conducts and radiates heat in the upstream direction;upstream of the reaction zone, the temperature of the solidexceeds that of the gas, and so there is solid-to-gas convectiveheat transfer. The incoming gases are thus preheated until theyreach the ignition temperature, reaction takes place, and the cyclecontinues.
s in a porous burner [7].
Fig. 3. (a) Schematic representation of heat recirculation in a porous medium
idealised as an insulated refractory tube and (b) the corresponding variation in gas
and tube wall temperature with distance [14].
ARTICLE IN PRESS
Fig. 5. Predicted peak flame temperatures for combustion in a porous medium in
comparison to a freely propagating, adiabatic, laminar flame [52].
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684670
The resulting temperature profiles for the solid and gas areshown in Figs. 3b and 4. The profiles shown in Fig. 4 wereobtained using a one-dimensional model containing full chem-istry and accounting for the effects of both solid and gasconduction, solid-to-solid radiation and convective heat transferbetween solid and gas [48]. The preheating effect is clearlyindicated, and it can be seen that although the reaction zoneitself—indicated by the peak in the heat release rate—is similar inwidth to that for a premixed laminar flame, the preheatzone—defined as beginning where the gas temperature hasincreased by 1% of its original inlet value and ending where thegas and solid temperatures are equal—is much wider.
Heat recirculation efficiency can be defined as the amount ofsolid-to-gas convection in the preheat zone compared to the firingrate. Based on this definition, Barra and Ellzey [48] predictedefficiencies of up to 25% (for the particular burner configurationmodelled). Additionally, it was found that the heat recirculationefficiency increases as the equivalence ratio (f) decreases. Therelative contributions of conduction and radiation to the recircu-lation process were also examined. It was found that at the lowestequivalence ratio investigated (f ¼ 0.55), conduction is moreimportant at lower flow velocities and vice versa at highervelocities. As the equivalence ratio (and hence temperature)increases, radiation becomes the dominant mechanism.
These findings corroborate those of an earlier study by Min andShin [49] using a one-dimensional model that considered all threemodes of heat transfer but modelled the combustion as a single-step irreversible reaction. It predicted that 28% of the total heatreleased would be recirculated to the incoming reactants, and thatthe contributions of conduction and radiation were approximatelyequivalent in magnitude (at f ¼ 0.55).
The degree of heat recirculation, and the role played by each ofthe heat transfer mechanisms, will obviously be highly dependenton the properties of the actual porous material being considered.The two previous examples modelled burners made of partiallystabilised zirconia (PSZ) foam and a honeycomb ceramic respec-tively. The influence of porous material is discussed in more detailin Section 4.1.
If a porous material that promotes effective heat recirculationis selected, excess enthalpy or super-adiabatic combustion in theporous medium can be realised. Peak temperatures greater thanthe adiabatic flame temperature, as well as flame speeds higherthan the associated laminar flame speed for the mixture, havebeen observed or predicted by several researchers [48,50–56].
For example, Figs. 5 and 6 show the predictions of Hsu et al.[52] obtained using a one-dimensional model of a PSZ burnerincluding full chemistry, separate energy equations for the gas andsolid phases, and radiative, conductive and convective heat
Fig. 4. Illustration of preheating in a porous burner demonstrating the existence of
an enlarged preheat zone [48].
transfer. From Fig. 5, which shows peak temperatures, it can beseen that super-adiabatic combustion is predicted over a widerange of equivalence ratios, with the effect becoming morepronounced for leaner mixtures. Flame speeds in excess of thelaminar flame speed are also calculated (Fig. 6). In addition, thepresence of the porous medium is predicted to extend the leanflammability limit (to f ¼ 0.36 for this particular case).
2.2.2. Flame stabilisation
In order to stabilise the combustion process within the porousmedium a balance must be achieved between heat recirculation,heat release and heat losses, such that the effective flame speed isequal to the incoming velocity. When the flow velocity is greaterthan the flame speed the flame will propagate downstream andvice versa (Section 2.2.3).
For the purposes of this review we are concerned only withcombustion that is actually stabilised within the porous medium(sometimes referred to as an ‘embedded’ or ‘submerged’ flame).The related phenomenon of surface combustion, whereby theflame is stabilised at or just above the surface of a porous bed, willnot be considered. This has been the subject of much previousresearch, for example [57–59]. The Combustion Technology groupat the University of Eindhoven have also published a useful seriesof studies on this topic [60–65].
It is difficult to predict a priori whether or not stablecombustion will be achieved in a particular porous burner for agiven fuel/air mixture, and, moreover, if stable combustion isachieved, at what position in the porous bed the flame willactually be located. Nonetheless, an intuitive explanation ofthe flame stabilisation process in general terms is offered byBuckmaster and Takeno [66]: when a change occurs in the inlet
ARTICLE IN PRESS
Fig. 6. Predicted burning speed and flammability limits for combustion in a
porous medium in comparison to a freely propagating, adiabatic, laminar flame
[52].
Fig. 7. Temperature profiles for a range of flow velocities at f ¼ 0.4 [69].
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684 671
conditions such that the flame moves downstream, for example ifthe flow velocity is increased, then if this movement causes anincrease in the flame speed, the flame will eventually reach alocation where the flame speed again matches the flow velocity,and it will stabilise in that new location. In other words, on a plotof flame speed against flame location, the flame can be stabilisedat those locations where there is a positive gradient.
The requisite increase in flame speed will occur due to theincoming gases being preheated more effectively. As a generalrule, in the upstream region of the porous bed the amount ofpreheating will increase as the flame moves downstream (as lessheat will be lost from the upstream end of the burner). It wouldtherefore be expected that stable combustion is likely to occur inthis region. This behaviour is predicted by a number of modellingstudies [45,67,68] and has been confirmed experimentally ininvestigations using a variety of burner configurations [47,49,51,55,69–72]. To give an example, Fig. 7 shows the resultsobtained by Afsharvahid et al. [69] using a burner with a porousbed of alumina spheres for a series of flow velocities at anequivalence ratio of 0.4.
Likewise, a decrease in fuel concentration is also predicted tocause the flame to move downstream [67,73]. Decreasing theequivalence ratio will result in a corresponding decrease in theflame speed. The flame must move to a new downstream locationin order for the flame speed to increase so that it once morematches the flow velocity. Again, this prediction is supported by anumber of experimental studies [47,55,69].
In this way, the heat capacity of the porous bed means that it ispossible to stabilise combustion over a range of flow rates and fuelconcentrations. Consequently, porous burners are characterisedby extended operating ranges and large turndown ratios. As an
additional means of stabilising the flame over a still wider rangeof conditions, a number of porous burners employ a two-sectiondesign whereby different porous materials are used in theupstream and downstream regions of the burner—a small-poredmaterial in the upstream, and a large-pored material in thedownstream section respectively—with the flame stabilising inthe downstream section at or near the interface. This concept isexamined in more detail in Section 4.2.
By the same principles, it might be expected that thecombustion process would be stable against short-term fluctua-tions in the flow and concentration. This was recognised in theanalysis of Takeno and Sato [45]: if the flow rate increasesmomentarily, the flame will move downstream—the resultingincrease in preheating will increase the flame speed and bring itback to its original position. This is a topic that has not yet beenexplored fully: most studies, both experimental and numerical,investigate only steady-state behaviour.
Henneke and Ellzey [74] did however examine the response ofporous burners to changes in fuel flow rate. Specifically, theyinvestigated the behaviour of the burner when the fuel supply wascompletely interrupted (although cold air continued to flowthrough the hot bed) and then re-introduced to determine if themixture would reignite under these circumstances. They used aone-dimensional model including full chemistry plus conductive,radiative and convective heat transfer and performed transientsimulations at an equivalence ratio of 0.7. They found that there isa ‘critical cooling time’ which is the longest time for whichreignition is still possible. If the fuel is re-introduced within thistime the mixture will reignite at the downstream end of theporous bed and the flame will then propagate upstream. Theirsimulations predicted a wide variation in critical cooling timesranging from approximately 10 to as much as 850 times theresidence time of the gas in the porous bed (between 1 and 69 sfor the system considered): it was found that porous materialswith lower porosities and higher heat capacities allowed the fuelsupply to be interrupted for longer.
2.2.3. Transient combustion
This review is concerned primarily with stationary combustionsystems, where for a given set of conditions (flow velocity andequivalence ratio) the flame is stabilised within the porousmedium, as described above. However, transient combustion(also commonly referred to as ‘filtration’ combustion) systems,based on a combustion wave propagating through the porousmedium, have also been examined in numerous studies (e.g.[75–94]). A great deal of research in this area was also carried outin Russia during the 1970s and 80s, and a selection of this earlywork is reviewed by Babkin and colleagues [95,96].
ARTICLE IN PRESS
Fig. 8. Stable operating range for a two-section FeCrAlY burner showing the
existence of both sub- and super-adiabatic modes of operation [56].
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684672
Researchers in the field of filtration combustion have identifieda number of distinct combustion regimes based on the speed atwhich the combustion wave propagates through the porousmedium. We are interested only in what is classified as the low-velocity regime, where the velocity of the combustion wave istypically less than 1 mm/s: it is only for this case that there issignificant heat transfer between the gas and solid [95].
As with a stable flame, this heat transfer will lead to heatbeing recirculated from the hot combustion products back to thecold reactants via the solid, resulting in super-adiabatic combus-tion. However, in transient combustion systems there is additionalcomplexity, because the displacement of the combustion zonerelative to the solid means that the super-adiabatic effect can beeither enhanced or inhibited, depending on the whether or notthe combustion wave is travelling with (‘co-flow’) or against(‘counter-flow’) the flow of the incoming gas: in the case of co-flow, there will be additional heat transfer from solid to gasupstream of the reaction zone, and the super-adiabatic effect willbe enhanced [96].
The direction of propagation of the combustion wave in a givensystem depends on both the flow velocity and the equivalenceratio. Under ultra-lean conditions, co-flow of the combustionwave is observed, leading to a pronounced super-adiabatic effect.For higher equivalence ratios approaching the normal lean limit(f ¼ 0.5), co-flow occurs at low flow velocities; as the flowvelocity is increased the velocity of the combustion wave alsoincreases through a maximum before falling to zero—in otherwords the flame is stabilised—before finally reversing directionand propagating against the flow of the incoming gas [85,96].
Transient combustion is significant to many lean-burn applica-tions and in particular offers advantages where the energy contentof the fuel is extremely low. Because the transient regime canpotentially provide more effective heat recirculation than astabilised flame, a greater extension of the lean flammabilitylimit should be possible. The challenge lies in exploiting thedesirable features of transient combustion in a practical systemwhere the combustion wave must be restricted to within theconfines of a burner.
2.2.4. Sub-adiabatic combustion
So far, we have been considering excess enthalpy or super-adiabatic combustion in a porous medium, as this is the means bywhich the lean flammability limit may be extended and ultra-leancombustion achieved. However, if the balance between heatrelease, recirculation and loss is altered such that the contributionof the recirculation component is decreased (and by implicationthat of heat loss increased), then a sub-adiabatic combustionregime may alternatively be observed. Heat ‘lost’ in this contextmeans heat that is not recirculated to preheat the incomingreactants: in a practical burner this heat may actually be usefullyrecovered from the system to heat some load, either via thethermal energy of the exhaust gases or radiant heating from theporous solid (Section 4.4).
It was shown previously that a flame might be expected tostabilise within the upstream half of a porous burner. However atlow velocities below the laminar flame speed, a second stableburning region at or near the downstream surface of the burner isalso predicted [49,50]. The regime that is observed will bestrongly dependent on burner design. Most studies examineeither one or the other of the sub- or super-adiabatic flame speedregimes. As a general observation, porous radiant burners, whichare designed to maximise radiant heating—that is heat ‘loss’—from the downstream surface, and which typically have arelatively thin porous bed of only a few centimetres in depth,and which operate on fuel/air mixtures closer to stoichiometric
(high temperatures being required to increase radiant output), arecharacterised by sub-adiabatic flame speeds [14,70,97]. However,it is also possible to observe both super- and sub-adiabatic flamespeeds in a single burner. Vogel and Ellzey [56] investigated atwo-section burner made of FeCrAlY metal foam. At equivalenceratios of 0.65 and below, stable combustion both above and belowthe laminar flame speed was observed; at higher equivalenceratios the burner could be operated in the sub-adiabatic modeonly, as illustrated by Fig. 8.
In the case of transient systems, sub-adiabatic combustion isobtained when the combustion wave propagates against thedirection of flow [85].
2.2.5. Emissions
A further consequence of the recirculation of heat away fromthe hot combustion products is that the subsequent decrease intemperature in this region inhibits the formation of nitrogenoxides (NOX). NOX may be formed either by the thermal(‘Zeldovich’) or prompt (‘Fenimore’) mechanism [98]. For condi-tions typical of porous burners, NO formed via the thermalroute—which is highly temperature dependent—constitutes themajority of NOX emissions [99]. Because NOX formation isdependent on the peak temperature, NOX emissions are observedto decrease with decreasing equivalence ratio [7,53,55,70,72,100].The effect of flow velocity on NOX emissions is less clear, but Durstand Trimis [7] and Khanna et al. [53] found that NOX emissions arerelatively insensitive to flow rate. This is attributed to the fact thatalthough reducing the flow velocity reduces the peak tempera-ture, this is compensated for by an increased residence time in theflame zone. It should be mentioned that because emissions aretypically very low (less than 30 ppm), any variations are likely tobe within the calculated error of the measurements [72], makingthe identification of clear trends difficult.
Conversely, reduced temperatures in the combustion zonemight be predicted to lead to increased emissions of the productsof incomplete combustion: carbon monoxide (CO) and unburnedhydrocarbons (UHC). However, the use of lean mixtures meansthat this is not the case. A number of investigations find that COemissions tend to be very low (less than 40 ppm) and that theydecrease with decreasing equivalence ratio but increase withincreasing flow velocity (as there will be less time for the COformed in the reaction zone to be oxidised to CO2 before exitingfrom the burner) [53,55,70,72]. Khanna et al. [53] also observedthat for very low flow velocities at equivalence ratios of 0.65 orless there is an increase in CO emissions, due to the lowtemperatures obtained under these conditions suppressing theoxidation of CO to CO2.
ARTICLE IN PRESS
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684 673
Emissions of UHC from porous burners also tend to benegligible (less than 5 ppm) and in so much as any dependencieson flow rate and fuel concentrations can be reliably identified theyfollow the same trends as for CO [55,70,72].
Finally, in order to minimise emissions, a homogeneoustemperature distribution within the burner is desirable. Hot spotsin the porous matrix can lead to an increase the formation of NOX,whereas cold spots might result in incomplete combustion and asubsequent increase in CO and UHC emissions [7].
Fig. 9. Stability diagram for combustion in a porous burner using of a bundle of
alumina tubes; flame type II represents combustion actually within the porous
medium [47].
3. Examples of burner performance
Much of the early experimental work on porous burnersfocused on their use as radiant burners (typically operating in thesub-adiabatic combustion regime), and consequently on measur-ing and optimising radiant output and efficiency. Howell et al. [8]review some of this work; see also [4,53,70] for some recentexamples. Radiant efficiencies reported cover a wide range ofvalues up to 40%. It is likely that the large variation in the reportedefficiencies is due to the lack of a standard procedure formeasuring radiant output [101] as well as to differences in burnerdesign and operating conditions.
We are more concerned with those studies that tell ussomething about the burner’s stable operating range. That is,over what range of flow rates and fuel concentrations can stableand complete combustion can be sustained? And in particular,what are the conditions—and how does the burner behave—at ornear the lean limit?
Outside of the stable operating range the flame will experienceeither blowoff (the flame propagates downstream because theinlet velocity is greater than the flame speed and eventually‘blows off’ the top of the burner), flashback (essentially theopposite situation), or, in the case of very lean mixtures, extinction
(because the temperature in the burner is not sufficient for theflame to sustain itself). In practice, some of these concepts are notso clearly defined: For example, blowoff can occur gradually, withthe flame starting to blow off one location on the burner surfacewhile still being stabilised at other locations [72]. Or, where multi-section burner designs (Section 4.2) are used, with a ‘preheating’section upstream of the main combustion section, the undesirableoccurrence of the flame stabilising within this preheating section(rather than leaving the burner altogether) is often consideredflashback.
Commonly the operating range of a porous burner will beillustrated by a ‘stability diagram’ or ‘burner map’ (e.g. Fig. 8),which plots either firing rate or velocity against equivalence ratio,showing the stable burning region as well as indicating thoseregions where blowoff, flashback or extinction occur. Alterna-tively, the burner’s performance may be more simply specified interms of its thermal power or load (kW), firing rate (kW/m2) orpower density (kW/m3).
As far as exploring burner operating range under leanconditions is concerned, there are two main bodies of work ofinterest: that carried out by Ellzey and colleagues at theUniversity of Texas at Austin [53,55,56,72,102,103], and thatundertaken by the combustion technology group at the Universityof Erlangen-Nuremburg (see [7,12] for an overview). The burnersused by the Austin group have all tended to be fairly similar, inthat they use a two-section design, combustion chambers ofdiameter 5–10 cm and reticulated ceramic (or more recentlymetallic) foams as the porous material. Those developed by theErlangen group have tested a wider variety of porous materialsincluding packed beds and lamella structures (Section 4.1.2) andhave been integrated with heat exchangers for energy recovery.
The performance of both sets of burner is fairly similarhowever, and can be summarised as:
(i)
Stable equivalence ratios ranging from 0.5 to fuel rich. (ii) Flow velocities of up to 2 m/s.
(iii)
Firing rates of up to 4000 kW/m2, but typically no more than3000 kW/m2. (Often the maximum firing rate may bedetermined not by the limitations of the burner itself, butby the practicalities of what can be measured in a laboratoryenvironment: for example, heating may become so great thatit becomes a hazard [55].)
(iv)
Emissions of NOX, CO, and UHC of less than 40 ppm (for leanmixtures).
(v)
Maximum pressure drops of 0.1 bar/m [72] (although this isnot usually reported, and will be highly dependent on theporous material used).
3.1. Examples of ultra-lean burner performance
Investigations involving ultra-lean combustion (below the leanflammability limit of f ¼ 0.5) are not at all common.
The early experimental work on combustion in a porousmedium by Kotani and Takeno [46] and Kotani et al. [47] resultedin combustion being maintained at equivalence ratios of about0.2, as shown by the burner map in Fig. 9. The burner used in thiswork combined the porous medium—in this case a bundle ofalumina tubes of inner diameter 0.6–0.8 mm—with an externalheat exchanger surrounding the combustion chamber.
In later work, Hsu et al. [51] carried out an experimental studyon a two-section burner design (Section 4.2) consisting of tworeticulated ceramic PSZ cylinders. The cylinders were each 5.1 cmin both diameter and length, and insulated circumferentially. Thepreheating section had a pore density of 25 ppcm and thecombustion section 4 ppcm. Fig. 10 shows the stability diagramobtained for this burner. It can be seen that there was a modestextension of the lean limit to f ¼ 0.41 and that above this limit
ARTICLE IN PRESS
Fig. 11. Temperature profiles within a porous bed of alumina spheres for various
equivalence ratios and firing rates (indicated by the number in brackets; in kW/
m2) [105].
Fig. 10. Stability diagram for a two-section PSZ burner [51].
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684674
stable combustion was possible over a wide range of velocities atany given equivalence ratio. These findings were corroborated by anumerical analysis of the same burner, which predicted the leanlimit to occur at f ¼ 0.43.
More recently, Christo and colleagues at the University ofAdelaide [69,104,105] investigated the performance of a largerburner consisting of an insulated cylinder of 15.4 cm diameter and60 cm in length filled with 6 mm diameter alumina spheres plus athin layer of flint clay beads at the base of the combustionchamber to act as a flashback arrestor. Only a limited selection ofpreliminary results were reported but these indicated that it waspossible to maintain combustion at f ¼ 0.1 at a firing rate of90 kW/m2 [105], as demonstrated by the steady-state temperatureprofiles shown in Fig. 11. However it can be seen that thesetemperature profiles are somewhat erratic, and as a complete setof results has not been published this finding should perhaps betreated with some scepticism as regards the stability andcompleteness of the combustion process: quite possibly this isactually an example of transient combustion, with the flamepropagating extremely slowly—at less than 1 mm/s—up theporous bed (Section 2.2.3). Nevertheless, emissions of CO andNOX were measured at equivalence ratios of 0.35 and 0.4 and werefound to be in the range 0–5 ppm in each case [69], indicating that
at these slightly higher fuel concentrations at least, combustionwas complete.
Finally, a number of studies on transient combustion systemshave shown unequivocally that it is possible to achieve asignificant extension of the lean limit into the ultra-lean regime.For example, Bingue et al. [78], using a 45 cm long, 3.8 cmdiameter combustion chamber filled with 3 mm alumina spheres,obtained combustion at equivalence ratios down to 0.25. At thisequivalence ratio, and at a flow velocity of 0.12 m/s, thecombustion wave propagated downstream with a velocity of0.4 mm/s. Kennedy et al. [90] operated a burner of the samedesign at a higher flow velocity of 0.25 m/s and were able toextend the lean limit to f ¼ 0.2.
So, from what has been reported in the literature, it can beconcluded that stable ultra-lean combustion in porous burners ispossible, but that there is a lack of reliable data relatingspecifically to this phenomenon. In part this is because themajority of researchers have hitherto been more concerned withother aspects of burner performance—such as radiant output—as previously discussed. However it is also indicative of a moregeneral lack of reliable measurements from porous burners—suchdata being inherently difficult to obtain due to the restrictivepresence of the porous solid. It might be expected that accuratein-pore measurements of temperature, species concentration, gasvelocity, turbulence intensity and so on would greatly enhanceour understanding of the processes at work.
4. Burner design considerations
As we have seen, combustion in a porous burner involvesstabilising a flame within the pores of a solid matrix, and entailsan intimate coupling of combustion, heat transfer and fluiddynamics. Despite our incomplete understanding of all theprocesses at work, we can still attempt to exploit the resultingdesirable performance characteristics—namely extended flamm-ability limits, high flame speeds and low emissions—in the designof practical burners for lean-burn applications.
Selection of a porous material with the correct thermo-physical properties is likely to be the single most importantdesign decision. Other factors to consider include the length of theporous bed, the use of multi-section designs, the shape andorientation of the combustion chamber, the integration of theburner with some form of external heat recirculation to provideadditional preheating of the incoming fuel/air mixture, and if (andhow) useful energy is to be extracted from the system.
4.1. Porous material selection
The principle of heat recirculation via a porous solid isfundamental to porous burner operation: choosing a porousmaterial with heat transfer properties that allow this recirculationto proceed effectively is therefore of primary importance.Conductive, convective and radiative heat transfer all contributeto the heat recirculation process, and so the thermal conductivity,convective heat transfer coefficient, emissivity and optical thick-ness (or its inverse, the radiative extinction coefficient) of amaterial must all be considered [106].
A number of studies have used models of varying degrees ofcomplexity to investigate the influence of these various heattransfer parameters on burner performance. They confirm thatincreasing the thermal conductivity or convective heat transfercoefficient increases the degree of heat recirculation[52,99,107,108].
For example, Barra et al. [107] used a one-dimensional modelincorporating a detailed chemical mechanism and separate
ARTICLE IN PRESS
Table 1Properties of some common ceramics [109]
Property Al2O3 SiC ZrO2
Maximum usage temperature in air
(1C)
1900 1600 2300
Thermal conductivity at 1000 1C
(W m�1 K�1)
5–6 20–50 2–4
Total emissivity at 2000 K 0.28 0.9 0.31
Coefficient of linear thermal
expansion from 20–1000 1C
(10�6 K�1)
8 4–5 10–13
Resistance to mild thermal shock
(10�3 W�1)
3 23 1
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684 675
energy equations for the gas and solid phases to investigate theinfluence of the thermal conductivity, convective heat transfercoefficient and radiative extinction coefficient on the stable rangeof flow velocities for an equivalence ratio of 0.65. They found thata high thermal conductivity and heat transfer coefficient result inboth the largest stable operating range and the highest maximumvelocity. Additionally, they discovered that there is an optimumvalue for the radiative extinction coefficient. They explained thisfinding as follows: If the extinction coefficient is too large thenradiation takes place over too small a distance; conversely, if theextinction coefficient is too small then the radiation will be spreadover too large an area; in either case the incoming gases will notbe preheated effectively.
However, all of the relevant studies to date have investigatedthe effect of changing the various heat transfer parameters at asingle equivalence ratio. A parametric study on the effect of eachof these parameters on the achievable lean limit has not beenreported.
Another important consideration when selecting a porousmaterial is the pore size (or its inverse, the pore density). This willinfluence the combustion process via its effect on heat transport.Typically, small-pored materials exhibit good conductive (moresolid contact surfaces) and convective (larger internal surfacearea) but poor radiative (low optical thickness) heat transfer andvice versa. There is clearly a payoff between the three differentmodes of heat transfer suggesting that for any given system therewill be an optimum pore size. Furthermore, additional heattransfer may occur due to thermal dispersion effects—that is,enhanced heat transfer due to hydrodynamic mixing of the gaswithin the pores—and these will also be strongly influenced bypore size and geometry [106].
Hsu et al. [51] investigated the effect of pore size on the stableoperating range of a two-section burner consisting of reticulatedfoam cylinders made of PSZ. Foams of three different poredensities (4, 12 and 18 ppcm) were tested in the combustionzone. The 4 ppcm foam—the largest pore size—was found to bethe most effective of the three at extending the lean limit,although of course whether or not the actual optimum pore size islarger or smaller than this was not revealed by this experiment.
The volumetric porosity of the material is also important.A high porosity and permeability are desirable in order tominimise the pressure drop across the burner, as well as todecrease the time required to preheat the porous bed during thestart-up phase [3].
The durability of the material must also be considered. Thereare certain obvious constraints that any material must satisfy,namely an application temperature above the burner operatingtemperature and resistance to oxidative or reductive atmospheres.The importance of a material’s ability to withstand the hightemperature gradients expected during burner operation, parti-cularly during start-up and shut-down, has also been noted [7,8]:a low thermal expansion and high resistance to thermal shock aretherefore desirable attributes.
Other general considerations include the cost and availabilityof the material, the convenience with which it can be employedand the ability to manufacture it in the desired geometry.
The overall performance of a porous material will depend onits particular combination of base material and porous structure.The solid material chosen will influence the overall properties ofthe porous matrix via its thermal conductivity, emissivity,temperature and corrosion resistance, thermal expansion, andmechanical strength at high temperatures. The geometricalstructure employed will affect the radiative heat transport viaits optical thickness, the conductive heat transport via theexistence and extent of contact surfaces or solid material bridges,and the convective heat transport via the porosity and pore size
and the resultant internal surface area and flow patterns. Theporous structure will also determine the pressure drop and have adeciding influence on the strength and thermal shock resistanceof the final material.
A wide variety of porous media have been used or suggestedfor porous burner applications. Useful reviews on this topic havebeen published by the combustion technology group at theUniversity of Erlangen-Nuremburg [3,7,109,110], as well as byHowell et al. [8], who concentrate on ceramic foams. The mainbase materials used are ceramics such as alumina, silicon carbideand zirconia, or high temperature metal alloys. Possible geome-trical structures include reticulated foams, packed beds andlamella structures.
4.1.1. Base materials
Ceramics: Ceramics are suitable for porous burner applicationsbecause of their high usage temperatures, chemical stability andresistance to erosion and wear [111]. The most commonly usedhigh temperature ceramics are alumina (Al2O3), silicon carbide(SiC) and zirconia (ZrO2), the relevant properties of which aresummarised in Table 1.
Alumina is the most popular, employed either in a packed bed[12,69,105,112] or as a lamella structure [15,100,113–115]. It has ahigh application temperature and is resistant to wear andcorrosion, as well as being economical. It has a moderate thermalconductivity and emissivity, but a large coefficient of thermalexpansion and poor thermal shock resistance [3,109,116]. Theproperties of any alumina-based ceramic depend on the actualalumina content: ceramics with higher silica contents willtypically have lower maximum usage temperatures and thermalconductivities [109].
SiC and SiSiC ceramics oxidise at around 600 1C [109], but aslong as the resulting surface layer of silica remains stable, can beused up to reasonably high temperatures. Compared to alumina,they have the benefits of a high thermal conductivity andemissivity, a lower coefficient of thermal expansion and verygood thermal shock resistance [3,109,116]. They are mostcommonly used as reticulated foams [12,15,17,24,38,100,115,117–119], but have also been employed in packed beds [112] andas static mixer structures [113].
Zirconia-based ceramics on the other hand generally have avery high application temperature but a low thermal conductivity,high coefficient of thermal expansion, and moderate thermalshock resistance and emissivity [3,109,116]. Pure zirconia under-goes a destructive phase change from tetragonal to monoclinicwhen cooled from the sintering temperature, and must bestabilised against this by the use of additives such as magnesia,yttria, calcium oxide or ceria [120,121]. The resulting stabilised orpartially stabilised zirconia will typically have a lower applicationtemperature—around 1800 1C—than the pure solid [109].Although the heat transfer properties of zirconia do not seem
usuario
Resaltar
ARTICLE IN PRESS
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684676
particularly favourable, various researchers have used zirconia-based ceramic foams in porous burners. Examples include PSZwith magnesia as the binder [51,53], yttria- and magnesia-stabilised zirconia [24], and an yttria-stabilised zirconia/aluminacomposite (YZA) [55,72]. The use of zirconia-toughened ceramicssuch as zirconia-toughened mullite (ZTM) [55] has also beenreported; another such possibility might be zirconia-toughenedalumina (ZTA) [121].
Other high temperature ceramics previously employed inporous burners include lithium aluminium silicate (LAS) [71],cordierite [70] and mullite [122].
Metals: Surface burners—where the flame is stabilised at thedownstream surface of the porous bed, rather than within theporous matrix itself—commonly use a mat composed of wovenmetal fibres as the porous material. In the past stainless steel wasusually used (e.g. [59]) and more recently high temperature alloyssuch as FeCrAlY have been employed [123]. Howell et al. [8]suggest that wire meshes are unsuitable for use in burners wherethe flame is actually located within the porous matrix, as thetemperatures encountered will lead to the rapid degradation ofsuch fine metal structures. Having said this, Huang et al. [124]reported stable combustion within a porous bed composed ofbundles of stainless steel wire mesh; no mention was made of thelong-term durability of the burner.
The development of high temperature alloys, combined with theability to produce them in reticulated foam structures [125],suggests that metals may become an increasingly valid andattractive alternative to ceramics for porous burner applications.Alloys such as FeCrAlY are designed for oxidation resistance andhigh temperature use, up to 1400 1C [126,127]. To date there is onlya single reported case [56] of the use of FeCrAlY foam in a porousburner, and the authors do not comment on the performance of thismaterial compared to its ceramic counterparts.
Carbon–carbon composites: Carbon–carbon (C–C) compositematerials (carbon fibre reinforcement in a graphite matrix) areanother alternative. C–C composites have excellent thermal shockresistance and thermal conductivity, however because theyoxidise readily at temperatures above 500 1C, a protective coating,for example SiC, rhenium, hafnium or iridium, must be applied. Aporous burner composed of a C–C composite foam coated in SiChas recently been developed [128].
4.1.2. Porous structures
Foams: Reticulated ceramic or metallic foams have an open-pore structure made up of an interconnected network ofdodecahedral-like cells, with solid struts forming the cell edges[129]. The literature contains numerous studies of porous burnersmade of reticulated ceramic foams, as documented above; the useof metallic foams is less common.
Reticulated foams are usually described in terms of avolumetric porosity plus linear pore density (number of poresper centimetre (ppcm)). Foams used in porous burners typicallyhave porosities in the range 70–90%, and pore densities in therange 2–25 ppcm, equating to nominal pore sizes (1/ppcm) in therange 0.4–5 mm, although actual pore sizes when measured aregenerally smaller—half the nominal size in some cases [8].
Foams typically exhibit good convective heat transport due totheir large internal surface area. Radiative and conductive heattransport are also generally good, with the optical thicknesstypically of the order of 10 pore diameters. Their high porositymeans that the pressure drop will be relatively low. They are alsolow in weight. Foams have the additional advantage of beingmanufacturable in a variety of complex shapes, and their rigidstructure also leads to flexibility in the angle at which the burnercan be operated [7,8,109].
However, this rigidity also leads to concerns about the durabilityof reticulated foams in a burner environment. Ceramic foamsexhibit poor thermal shock resistance and low fracture toughness.Hsu et al. [51] described the deterioration of a burner made of PSZfoam: the propagation of cracks though the foam eventuallydisturbed the uniformity of the flow causing the flame to tilt.
Elverum et al. [130] investigated the durability of YZA foams ina porous burner. They performed compression tests on foamsamples before and after testing in a typical burner environment.Degradation in compressive strength as a result of the burner testswas observed, with most of the damage found to occur due to thehigh thermal gradients present during the burner start-up phase.A scanning electron microscope (SEM) analysis revealed thatfailure involved the propagation of cracks from intrastrutpores—defects resulting from the manufacturing process.
Schmidt et al. [119,131] also performed durability tests, in thiscase on SiSiC foams. They found that the silica layer that forms onthe surface—and inside the intrastrut pores—of such foams,whilst stable under steady-state conditions, is quickly destroyedunder thermal cycling (in other words after repeated start-up andshut-down). The formation of cracks in the foams was attributedto tensile stresses resulting from a thermal mismatch between theSiSiC foam itself and the silica coating. Attempts to improve thedurability of the foams by applying a cordierite-based coating toprotect against corrosion were unsuccessful.
Material durability is especially important if porous burnersare to be successfully commercialised. To this end, researchers atthe University of Erlangen-Nuremburg have instigated a projectinvolving long-term (up to 3000 h of thermal cycling) durabilitytests, which aims to determine the likely performance of anumber of different ceramic foams over a burner’s lifetime; thiswork is ongoing [132].
Packed beds: Packed beds of discrete particles are a commonlyused alternative to foams. They have the advantage of increaseddurability as the particles are small robust shapes, and are notconstrained in a rigid matrix. Packed beds of ceramic spheres,typically alumina and of diameter 5–20 mm, are commonly used inporous burners (Section 4.1.1). However, packed beds of spheres arecharacterised by relatively low porosities—in the range 30–50%depending on bead size [8]—and the ensuing disadvantages. Theporosity can be increased by using an irregular packing shape suchas a Raschig ring or saddle. Using a packed bed of saddles, forexample, can increase the porosity to 90% or above [16]. Variousbespoke packing shapes (e.g. Fig. 12) intended for VOC destructionin regenerative thermal oxidisers are also available [133]. Thesehave been designed in a shape that provides optimal heat transferand so should be suitable for use in porous burners. Xiong et al.[112] studied a burner comprising a packed bed of ‘irregularlyshaped’ SiC particles, but otherwise the use of packing shapes otherthan spheres has not been explored to any great extent.
Lamella structures: Structures, such as those found in staticmixers, composed of several perforated ceramic lamellas arrangedside by side and twisted with respect to one another (Fig. 13) areanother alternative, and the use of lamellas made out of aluminafibres has been reported [113–115]. These structures have a veryhigh porosity of over 95%, and consequently a very low pressuredrop and short start-up phase. They have a high internal surfacearea and so good convective heat transfer. The open structuremeans that radiative heat transfer is very high, although theconductive heat transport is negligible. They also reportedly exhibitgood mechanical stability and thermal shock resistance [3,109].
4.1.3. Porous material selection for lean-burn applications
The most appropriate porous material will depend on theapplication for which the burner is intended. With respect to
ARTICLE IN PRESS
Fig. 13. Lamella structure made of alumina fibres [3].
Fig. 12. Random packing shape for use in regenerative thermal oxidisers [133].
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684 677
ultra-lean methane combustion, maximising the heat recircula-tion assumes primary importance. It has been shown that at lowerequivalence ratios conduction rather than radiation becomes thedominant mode of heat transport [48], so a material with asuperior thermal conductivity would be favourable. Furthermore,when operating on very lean mixtures, maximum burnertemperatures will be fairly low at around 1200 1C or less [105]:this means one’s choice of porous material is extended to includematerials with only a moderate maximum usage temperature.Given all of the above, foams made of either SiC or a metal alloyseem most promising.
Also, in order to improve heat recirculation we are interestedparticularly in maximising the axial heat transfer against thedirection of flow. With the exception of lamella structures, whichare probably not suitable for lean-burn applications due to theirlow thermal conductivity, the materials discussed so far have beenhomogeneous, transporting heat uniformly in all directions.However materials capable of transferring heat preferentially inthe desired direction could potentially be superior, for examplethe bundle of alumina tubes used in the experiments of Kotaniand Takeno [46] and Kotani et al. [47], or the honeycomb structureemployed by Min and Shin [49]. This is certainly an area thatwarrants further research.
As previously described, a number of modelling studies haveinvestigated the effect of altering each of the various heat transferparameters on burner operation. In a real material the effectsof the different parameters cannot be isolated in this way(for example decreasing the pore size might increase conductionbut it will also decrease radiation). Also, the relevant thermo-physical properties of many common materials, particularlyreticulated foams, are poorly characterised. This hinders thedevelopment of accurate models, as some of the most importantparameters will not be reliably known. It is consequently difficultto carry out modelling studies that compare actual availablematerials.
Moreover, there are few experimental investigations thatobjectively compare the performance of one or more porousmaterial in the same burner (some studies that do attempt to dothis are described below). It is difficult to analyse the performanceof porous materials across different studies because the effect ofthe porous material cannot be isolated from other differences inburner design or experimental procedure.
Xiong et al. [112] compared the performance of a porousburner (with an integrated heat exchanger) made of a packedbed of alumina spheres to one composed of irregularly shapedSiC particles. They observed that combustion could be stabi-lised over a wider range of fuel concentrations for the SiC burner,indicating that the SiC provided more effective heat recircula-tion. They also investigated the effect of porous material onemissions. NOX emissions were found to be lower but COemissions higher in the SiC bed because of the more effectivecooling of the post-flame region. However, not enough informa-tion was provided about the two materials (for example particlesize was not given) to be able to draw useful conclusions fromthis study.
Mathis and Ellzey [55] investigated two different materials—
YZA and ZTM reticulated foams—in a two-section burner, withthe aim being to stabilise combustion at the interface betweensections. The foams were identical, other than in their use ofdifferent base materials. It was found that the YZA burnerstabilised combustion over a range of firing rates at anequivalence ratio of 0.65, but that the ZTM burner did notstabilise the flame effectively: at lower firing rates the flamepropagated into the upstream preheating section. Incompleteknowledge of the properties of the two different materialsmeans it is not possible to determine precisely why this wasthe case.
Al-Hamamre et al. [15] compared two different burners—eachwith the same dimensions—for the combustion of landfill gas(CH4 in CO2) and low-calorific value syngas (a mixture of CO, CH4
and H2 in CO2 and N2). The first burner was composed of aluminalamellas and the second of SiC foam: they found that for each ofthe fuels investigated the SiC burner was more effective atextending the lean limit.
These studies, although limited, highlight the importance ofcorrect material selection on burner operation, and the need forfurther research in this area.
usuario
Resaltar
usuario
Resaltar
usuario
Resaltar
usuario
Resaltar
ARTICLE IN PRESS
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684678
4.2. Multi-section design
A two-section or ‘bi-layered’ design consisting of a small-poredupstream section in which the incoming gas mixture is preheated,and a large-pored downstream section in which the combustionprocess actually takes place was patented by Durst and Trimis ofthe University of Erlangen-Nuremburg in 1993 [134], and most ofthe recent literature describes porous burners constructedaccording to this design [4,51,53,55,56,69,70,72,100,102,105]. Atwo-section design performs one or both of the following roles:first, the interface between the two sections provides anadditional means of stabilising the combustion process over awide range of flow rates and second, the upstream section acts asa flashback arrestor or ‘flame support layer’.
It has been suggested that the interface between the twosections acts as a ‘flameholder’, with the sudden change in poresize at the interface leading to local quenching of the flame [7].Chaffin et al. [102] sought to enhance the role of the interface asflameholder by actually inserting a water-cooled brass ringaround the circumference of the burner at that location. Theyasserted that this design allowed the stability range to befurther extended, although no results comparing burner perfor-mance with and without the addition of this device werepresented.
An alternative explanation for the role of the interface incombustion stability is suggested by Hsu et al. [51] based on theexplanation for flame stability offered by Buckmaster and Takeno[66] (Section 2.2.2) whereby on a plot of flame location againstflame speed, the flame can be stabilised at those locations wherethere is a positive gradient. For a two-section design, there is alarge jump in flame speed across the interface due to the increasein pore size, resulting in the ability to stabilise combustion over alarger flow range.
The upstream section can additionally serve as a flashbackarrestor if the pore size in this section is chosen to be less than theminimum required for flame propagation. Materials with poresizes of the order of 1 mm or less are typically used.
The two-section design is not always employed to optimaleffect in its capacity either as flameholder or flashback arrestor.Some investigations have shown that the flame will propagateaway from the interface when the flow velocity is increased or theequivalence ratio decreased, with the combustion process even-tually stabilising in some downstream location [55,69,105].Conversely, there are reports of the flame propagating into, andsometimes stabilising within, the preheating region [51,55].
Previously (Section 4.1), the properties desirable in the porousmaterial used in the main combustion region of the burner weresurveyed. The sought-after properties for the material used in thepreheating section will be slightly different, most obviously therequirement for a smaller pore size as already discussed.Additionally, poor heat transport properties might make for amore effective flashback arrestor. There seems to be a conflict herethough, because if the flame is to be stabilised at the interface,the upstream section must also serve to preheat the incominggases (although in practice some degree of preheating will alsooccur at the start of the downstream combustion section),implying that the convective heat transport properties at leastshould be good.
The findings of Barra et al. [107] support the first view. Theymodelled the effects of the heat transport properties of theupstream section in a two-section burner at an equivalence ratioof 0.65. They found that a low thermal conductivity andconvective heat transfer coefficient resulted in the largest stableoperating range. They also concluded that it was the properties ofthe upstream section that significantly determined the minimumachievable flow velocity.
Examples of materials that have been used in the preheatingsection include packed beds of flint clay beads [69,105] andreticulated foams composed of PSZ [51,53,103], YZA [55] andFeCrAlY [56].
The length of the preheating section might also be expected toinfluence burner performance. Mathis and Ellzey [55] tested twodifferent preheating sections of length 2.5 and 5.1 cm respectivelyin a two-section YZA burner. The length of the main combustionsection was 5.1 cm and the pore densities of the foam were 23.6and 3.9 ppcm in the upstream and downstream sections respec-tively. They reported that the longer upstream section allowed theminimum stable firing rate to be extended, although the effectwas not found to be significant.
The use of three-section burner designs has also been reported.Brenner et al. [113] used a burner comprising a preheat sectionmade of 18 ppcm zirconia foam, followed by a section of 8 ppcmzirconia foam, followed by lamellas made of alumina fibres; thebenefit of using such an arrangement was not elucidated however.Hsu [135] also reported using a three-section design, again in azirconia foam burner. In this case the three sections formed a‘sandwich’ structure, with small-pored sections both upstreamand downstream of the main combustion region. The third sectionwas intended to provide both a means of redirecting radiative fluxback into the combustion zone and an additional interface atwhich the flame could stabilise. It was shown that this three-section design supported a broader range of flow velocities,although the lowest equivalence ratio investigated was only 0.7.
Finally, it should be noted that the use of multi-section designsinvolving regions of porous material with smaller pores or lowerporosities will increase the pressure drop across the burner, andthis must be accounted for when assessing the relative merits of agiven design.
4.3. Shape and orientation of combustion chamber
The default configuration for a porous burner is a cylindricalcombustion chamber oriented vertically such that the incomingfuel/air mixture flows upwards through the porous bed; themajority of reported studies use this construction. Square orrectangular cross-sectional geometries are also fairly common(e.g. [97,105,112,113]), and more so for porous radiant burners.Trimis and Durst [12] suggest that for larger burners, for someapplications a rectangular design might be preferable, as onedimension could then be kept reasonably small. This might bedesirable if the combustion chamber were surrounded by a heatexchanger for example. Moßbauer et al. [3] reported a porousburner with a ring-shaped cross-section, in the centre of which aconventional premixed burner could be operated, potentiallyallowing a further extension of the operating range. Althoughcertain configurations might be appropriate for particular appli-cations, it seems unlikely that the shape of the cross-sectionwould have much effect on the combustion process itself.
As regards the orientation of the combustion chamber, Huanget al. [124] used a horizontally oriented burner. A number ofstudies by the combustion technology group at the University ofErlangen-Nuremburg used a vertically oriented chamber with thegas flowing downwards. This arrangement was used because theirburner incorporates heat exchanger tubes in the porous beddownstream of the combustion region; a downward-flow ar-rangement allows the condensed water thus produced to flow outof the base of the burner. As before, the orientation of the burneris unlikely to have a significant effect on the combustion processper se.
All the burner designs considered so far have essentiallyinvolved the flow of gas through a combustion chamber in the
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684 679
axial direction (Fig. 14a). However cylindrical (Fig. 14b) or evenspherical (Fig. 14c) geometries where the flow is in the radialdirection have also been suggested. Such geometries have beenanalysed numerically [10,136–138] and it has been shown thatbecause the flow velocity is inversely proportional to radiallocation they should provide a natural means of stabilising theflame. A porous surface burner operating on natural gas/hydrogenmixtures using the cylindrical radial-flow arrangement wasreported by Brockerhoff and Emonts [139]. More recently Kamaland Mohamad [114, 140] used a similar arrangement to study theenhancement of the combustion of methane/air mixtures by swirlimparted by rotation of the central burner tube. Otherwise therehave been no practical realisations of these geometries.
Finally, an interesting modification to the normal cylindricalaxial-flow burner design is suggested by Durst and Trimis [7]. Ifthe radius of the combustion chamber at the start of thecombustion region increases gradually to form a funnel shape,then at lower firing rates only the smaller cross-sections are used,allowing a sufficiently high velocity to be maintained. Theopposite is true at higher firing rates and in this way the stableoperating range of the burner can be enlarged.
4.4. Recovery of useful energy
The two main means by which useful energy can be extractedfrom a porous burner are the use of the burner for radiant heatingor its integration with a heat exchanger to allow the recovery ofthermal energy. Viskanta [141] provides an overview andcomparison of these two approaches.
Radiant heating has been the focus of much of the previousresearch on porous burners—see Howell et al. [8] for a review ofthe topic. Porous radiant burners work on the principle that if asuitably emitting porous medium is used, the energy in the fuelcan be converted to radiant energy and used to heat a load. Porousradiant burners must therefore be designed such that theavailable radiating surface—usually the downstream end of theburner, but it could be the circumferential surface [103]—ismaximised, and the flame is stabilised close to this surface. Theseburners therefore usually have a relatively thin porous bed.Because a high temperature is required to maximise the radiantoutput, they are also typically run on close to stoichiometricmixtures.
Recovery of thermal energy by means of some form ofintegrated heat exchanger is less well documented. However anumber of investigations have looked at inserting heat exchangertubes into the porous matrix, either within, or downstream of, the
combustion zone [7,26,100,112,142,143]. The working fluid isheated by the hot combustion products as well as by radiationfrom the porous bed itself. The presence of cold surfaces in theflame zone is to be avoided, as this could lead to incompletecombustion and consequently increased CO and UHC emissions[12], so designs where the heat exchange takes place downstreamof the actual combustion region are preferable.
In addition, the incorporation of porous media into thecombustion chambers of gas turbines [3] or internal combustionengines [144,145] has been suggested. The direct generation ofelectricity in thermoelectric [146] or thermophotovoltaic[147,148] systems based around combustion in a porous mediumhas also been advocated.
4.5. Provision of supplementary external preheating
As we have seen, heat exchangers can be integrated withporous burners as a means of extracting useful energy from thecombustion process. However it is also possible to use an externalheat exchanger to recover and recirculate heat from the exhaustgases to preheat the incoming fuel/air mixture, in order tosupplement the internal heat recirculation provided by the porousbed.
This concept was demonstrated in early experimental work onporous burners carried out in the early 1980s by Takeno andcolleagues [46,47]. Their burner used a bundle of alumina tubes asthe porous medium and combined this with external heatrecirculation as shown schematically in Fig. 15. They were ableto maintain stable combustion at equivalence ratios of 0.2 usingthis design.
Around the same time Echigo and co-workers [142] at theTokyo Institute of Technology developed a porous burner whosecombustion chamber contained a ceramic plate burner followedby a layer of stainless steel mesh to facilitate heat recirculation.The steel mesh was surrounded by an arrangement of heatexchanger tubes for preheating the incoming combustion air.They operated the burner on a mixture of methane and hydrogenand were able to significantly extend the lean limit (from 0.284to 0.1).
More recently, Christo et al. [105] used a pilot-scale burnerincorporating preheating of the incoming reactants by the exhaustgases in an external heat exchanger to demonstrate the concept ofultra-lean combustion of LPG, however not much informationabout the configuration of this burner was provided.
The concept of providing supplementary external preheatingin this way has not been extensively investigated. Given that
ARTICLE IN PRESS
Fig. 15. Design of a porous burner incorporating external preheating [47].
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684680
in each of the above examples the lean flammability limitwas extended, this seems a highly promising area for furtherresearch.
Fig. 16. Temperature profiles in a reciprocal flow burner for various equivalence
ratios at a flow velocity of 20 cm/s [79].
4.6. Reciprocal flow operation
Previously (Section 2.2.3), we discussed the advantages oftransient combustion for ultra-lean applications, and it wasmentioned that the main challenge associated with developing apractical transient combustion system was finding a way toconfine a propagating flame within the limits of a stationaryburner. One way of doing this is to use a burner design featuringreciprocal flow operation.
In a reciprocal flow burner, the direction of flow is periodicallyreversed by means of flow switching values at either end of thecombustion chamber. For a combustion wave propagating in thesame direction as the incoming gas, if the length of each half-cycleis controlled so that it is greater than the residence time of the gas,but less than that of the combustion wave, then the combustionwave can be restricted to within the limits of the burner. Eachtime the flow direction is switched, the incoming gas is preheatedas it passes through the hot region of the porous bed where thecombustion and post-combustion zones were located in theprevious half-cycle [89].
In theory, reciprocal flow operation will result in a trapezoidaltemperature profile [84], although in practice due to radialheat losses an M-shaped profile is likely to be observed [79,86],as illustrated in Fig. 16. It can be seen that at lower equiva-lence ratios, the two temperature peaks are closer together, withthe profile eventually taking on a triangular shape at the leanlimit.
As with those operating with a stable flame, reciprocal flowburners may be integrated with heat exchangers—either em-bedded in the terminal sections of the combustion chamber [79],or surrounding it [87]—in order to extract energy or to providesupplementary preheating to the incoming combustion air, aspreviously discussed.
One of the main disadvantages of the reciprocal flow burnerdesign is the need to maintain reliable operation of themechanical flow switching values [82]. And in addition, a morecomplex burner control strategy will be required.
4.7. Other design considerations
Depth of porous bed: The required depth of the porous bed alsoneeds to be considered. As has previously been discussed, thebalance between heat release, heat recirculation and heat losseswill determine the position at which the flame stabilises withinthe porous bed, and changes in flow rate or fuel concentrationmay cause the flame to propagate up- or downstream beforestabilising at a new location. However it is difficult both to predicta priori at what location the flame will stabilise for a given burner,and to determine what the effect of changing the depth of theporous bed on burner performance will be.
Active flame stabilisation: The incorporation of some form ofwhat can be described as ‘active flame stabilisation’ [149,150] intothe burner design—and operating procedure—as a means ofstabilising and maintaining the flame at a desired location in theporous bed might also be considered. Essentially this involvesmonitoring the bed temperature and using a programmablecontrol system to automatically adjust the flow rate andcomposition of the incoming fuel air mixture—either by addingsupplementary fuel or additional dilution air—to counteract any
ARTICLE IN PRESS
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684 681
changes observed in the temperature profile. This capability couldbe especially valuable in ultra-lean combustion systems, as manylow-grade fuels are subject to unplanned fluctuations in both flowrate and methane concentration.
Insulation: In order for heat to be recirculated efficiently, radialheat losses should be minimised. Typically the combustionchamber will be encased in an insulating sleeve made of fibrousalumina or some similar material.
Flow distribution: For effective operation, a uniform flowdistribution is required, as well as thorough mixing of the fueland air feeds (if they are not premixed). The design of the burnerinlet will determine whether or not this is the case, so theprovision of a flow-straightening device such as a perforated plateor a honeycomb grate [26] might be considered.
Non-premixed operation: The operation of porous burners withnon-premixed flames is also possible [117,122]. However, becausethe presence of the porous material inhibits mixing of the fuel andair—increasing the likelihood of incomplete combustion—it isdesirable in this case to stabilise the flame just ahead of theporous matrix. This allows the fuel and air to mix before actuallyentering the solid, although the reaction zone may extend into theporous region itself and there will of course still be heat transferbetween the burnt gases and the solid matrix.
High-pressure operation: For applications where a highthroughput is required, operating the burner at elevated pressuresmight be an option. Moßbauer et al. [3] reported that a high-pressure porous burner chamber had been developed for use in agas turbine, but no results were presented. Noordally et al. [151]developed a burner—also intended for gas turbine applications—
with a nominal maximum operating pressure of 18 bar, althoughthe practical maximum operating pressure (at f ¼ 0.6) was foundto be around 12 bar, above which increasing temperatures in theinlet lead to concerns over flashback. And more recently,Altendorfner et al. [152] designed a burner intended to operateat elevated pressures of 10–15 bars, however so far onlyexperiments under atmospheric conditions have been reported.This is an area that needs further investigation.
EGR and staged operation: Other burner design featuresreported in the literature include the use of exhaust gasrecirculation (EGR) [153] and the staged addition of the reactants[103,115,154], however neither of these is particularly relevant tolean-burn applications.
5. Conclusions and recommendations for further research
Porous burners operate on the principle that when a premixedfuel/air mixture burns within the cavities of a solid porousmatrix, the solid serves as a means of recirculating heat from thehot combustion products to the incoming reactants, leading toexcess enthalpy burning. Combustion in a porous medium ischaracterised by increased flame speeds, extended flammabilitylimits, stability across a wide range of conditions, and lowemissions.
There is an extensive body of research relating to porousburners comprising both experimental and numerical investiga-tions. Despite this, some of the fundamental processes involvedare still not well understood, due both to the difficulties involvedin obtaining accurate experimental measurements from withinthe solid matrix and because some of the relevant properties ofthe most commonly used porous materials are not reliably known.This presents problems both for the development of accuratemodels, and for the design of burners to meet the particularrequirements of a given lean-burn application.
A number of specific topics have been identified that have notbeen explored fully to date and where further research would
therefore be beneficial. These include:
(i)
The effect of the heat transfer properties of the porousmaterial on burner operating range.
(ii)
The use of novel porous materials including alternativepacking shapes, metal foams, and materials capable oftransporting heat preferentially in one direction.
(iii)
A better characterisation of commonly used porous materials. (iv) An objective and comprehensive comparison of materials
and their effect on burner performance.
(v) The effect of porous bed depth on burner operating range.
(vi)
The effect of operating the burner at elevated pressures. (vii) The effect of the use of supplementary external preheating of
the incoming fuel/air mixture.
With respect to ultra-lean combustion in particular, althoughthe phenomenon has been observed in a handful of experimentalstudies, there is a lack of research focused specifically on thistopic. Further work—both experimental and modelling—is re-quired in this area generally, and as regards the optimisation ofburner designs for lean-burn applications particularly.
Additionally, there has been little attention paid thus far toissues regarding scale-up and how the technology might bepractically applied outside of the laboratory: this merits furtherconsideration.
Acknowledgements
We gratefully acknowledge the financial support of theAustralian Coal Association Research Program (ACARP), as wellas the School of Chemical and Biomolecular Engineering at theUniversity of Sydney for the award of a scholarship to Susie Wood.
References
[1] GoGas. Porous burner RADIMAX. Available from: /http://www.gogas.deS.Last accessed: 18 March 2008.
[2] Promeos. Porous burner. Available from: /http://www.promeos.comS. Lastaccessed: 18 March 2008.
[3] Moßbauer S, Pickenacker O, Trimis D. Application of the porous burnertechnology in energy- and heat-engineering, In: Proceedings of the fifthinternational conference on technologies combustion for a clean environ-ment (Clean Air V), Lisbon, 1999. p. 519–23.
[4] Rumminger MD. Numerical and experimental investigation of heat transferand pollutant formation in porous direct-fired radiant burners. PhD thesis,University of California: Berkeley; 1996.
[5] Von Issendorff F, Stamatov V, Mach A, Trimis D. Application of porousburners in industrial furnaces for glass production. In: Proceedings of theeighth international conference on technologies combustion for a cleanenvironment (Clean Air VIII). Lisbon; 2005.
[6] Trimis D, Pickenacker O, Wawrzinek K. Porous burners. In: Scheffler M,Colombo P, editors. Cellular ceramics: structure, manufacturing, propertiesand applications. Weinheim: Wiley; 2005. p. 484–508.
[7] Durst F, Trimis D. Combustion by free flames versus combustion reactors.Clean Air 2002;3:1–20.
[8] Howell JR, Hall MJ, Ellzey JL. Combustion of hydrocarbon fuels within porousinert media. Prog Energy Combust Sci 1996;22:121–45.
[9] Kamal MM, Mohamad AA. Combustion in porous media. P I Mech Eng A—
J Power 2006;220:487–508.[10] Mohamad AA. Combustion in porous media: fundamentals and applications.
In: Ingham DB, Pop I, editors. Transport phenomena in porous media, vol. 3.Pergamon Press; 2005.
[11] Oliveira AAM, Kaviany M. Nonequilibrium in the transport of heat andreactants in combustion in porous media. Prog Energy Combust Sci2001;27:523–45.
[12] Trimis D, Durst F. Combustion in a porous medium—advances andapplications. Combust Sci Technol 1996;121:153–68.
[13] Viskanta R. Combustion and heat transfer in inert porous media. In: Vafai K,editor. Handbook of porous media. 2nd ed. New York: Marcel Dekker; 2005.p. 607–44.
[14] Viskanta R, Gore JP. Overview of cellular ceramics based porous radiantburners for supporting combustion. Clean Air 2000;1:167–203.
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684682
[15] Al-Hamamre Z, Diezinger S, Talukdar P, von Issendorff F, Trimis D.Combustion of low calorific value gases from landfills and waste pyrolysisusing porous medium burner technology. Process Saf Environ 2006;84:297–308.
[16] Perry RH, Green DW. Perry’s chemical engineers’ handbook. 7th ed.New York: McGraw-Hill; 1997.
[17] Alavandi SK, Agrawal AK. Experimental study of combustion of hydrogen-syngas/methane fuel mixtures in a porous burner. Int J Hydrogen Energy2008;33:1407–15.
[18] Bakry A, Genenger B, Schmidt V, Trimis D. Low emission combustion ofvegetable oils with the porous burner technology. In: Afgan NH, CarvalhoMdG, editors. New and renewable technologies for sustainable development.Boston: Kluwer Academic Publishers; 2000. p. 541–55.
[19] Bingue JP, Saveliev AV, Fridman AA, Kennedy LA. Hydrogen sulfide filtrationcombustion: comparison of theory and experiment. Exp Therm Fluid Sci2002;26:409–15.
[20] Fuse T, Kobayashi N, Hasatani M. Combustion characteristics of ethanol in aporous ceramic burner and ignition improved by enhancement of liquid-fuelintrusion in the pore with ultrasonic irradiation. Exp Therm Fluid Sci2005;29:467–76.
[21] Gauthier S, Lebas E, Baillis D. SFGP 2007—natural gas/hydrogen mixturecombustion in a porous radiant burner. Int J Chem React Eng 2007;5:A114.
[22] Jugjai S, Polmart N. Enhancement of evaporation and combustion of liquidfuels through porous media. Exp Therm Fluid Sci 2003;27:901–9.
[24] Kaplan M, Hall MJ. Combustion of liquid fuels within a porous media radiantburner. Exp Therm Fluid Sci 1995;11:13–20.
[25] Kayal TK, Chakravarty M. Combustion of suspended fine solid fuel in airinside inert porous medium: a heat transfer analysis. Int J Heat Mass Trans2007;50:3359–65.
[26] Liu JF, Hsieh WH. Experimental investigation of combustion in porousheating burners. Combust Flame 2004;138:295–303.
[27] Pedersen-Mjaanes H, Chan L, Mastorakos E. Hydrogen production from richcombustion in porous media. Int J Hydrogen Energy 2005;30:579–92.
[28] Saveliev AV, Kennedy LA, Fridman AA, Puri IK. Structures of multiplecombustion waves formed under filtration of lean hydrogen–air mixtures ina packed bed. In: Proceedings of the twenty-sixth symposium (interna-tional) on combustion. The Combustion Institute; 1997. p. 3369–75.
[29] Tseng C-J, Howell JR. Combustion of liquid fuels in a porous radiant burner.Combust Sci Technol 1996;112:141–61.
[31] Kesting A, Pickenacker O, Trimis D, Durst F. Development of a radiationburner for methane and pure oxygen using the porous burner technology.In: Proceedings of the fifth international conference on technologiescombustion for a clean environment (Clean Air V). Lisbon; 1999.
[32] Bingue JP, Saveliev AV, Fridman AA, Kennedy LA. Hydrogen production inultra-rich filtration combustion of methane and hydrogen sulfide. Int JHydrogen Energy 2002;27:643–9.
[33] Bingue JP, Saveliev AV, Kennedy LA. Optimization of hydrogen production byfiltration combustion of methane by oxygen enrichment and depletion. Int JHydrogen Energy 2004;29:1365–70.
[34] Dhamrat RS, Ellzey JL. Numerical and experimental study of the conversionof methane to hydrogen in a porous media reactor. Combust Flame2006;144:698–709.
[35] Drayton MK, Saveliev AV, Kennedy LA, Fridman AA, Li Y. Syngas productionusing superadiabatic combustion of ultra-rich methane–air mixtures. In:Proceedings of the twenty-seventh symposium (international) on combus-tion. The Combustion Institute; 1998. p. 1361–7.
[36] Hall MJ, Peroutka XN. A porous media burner for reforming methanol forfuel cell powered electric vehicles. Society of Automotive Engineers SpecialPublications SP-1108; 1995. p. 101–9.
[37] Slimane RB, Lau FS, Khinkis M, Bingue JP, Saveliev AV, Kennedy LA.Conversion of hydrogen sulfide to hydrogen by superadiabatic partialoxidation: thermodynamic consideration. Int J Hydrogen Energy 2004;29:1471–7.
[38] Afsharvahid S, Ashman PJ, Dally BB. Investigation of NOX conversioncharacteristics in a porous medium. Combust Flame 2008;152:604–15.
[39] Glassman I. Combustion. 3rd ed. San Diego: Academic Press; 1996.[40] Weinberg FJ. Combustion temperatures: the future? Nature 1971;233:
239–41.[41] Hardesty DR, Weinberg FJ. Burners producing large excess enthalpies.
Combust Sci Technol 1974;8:201–14.[42] Zabetakis MG. Flammability characteristics of gases and vapors. In: Bureau
of mines bulletin 627. Washington: United States Bureau of Mines; 1965.[43] Jones AR, Lloyd SA, Weinberg FJ. Combustion in heat exchangers. Proc Roy
Soc Lond A Mat 1978;360:97–115.[44] Lloyd SA, Weinberg FJ. A burner for mixtures of very low heat content.
Nature 1974;251:47–9.[45] Takeno T, Sato K. An excess enthalpy flame theory. Combust Sci Technol
1979;20:73–84.[46] Kotani Y, Takeno T. An experimental study on stability and combustion
characteristics of an excess enthalpy flame. In: Proceedings of the Nine-teenth symposium (international) on combustion. The Combustion Insti-tute; 1982. p. 1503–9.
[47] Kotani Y, Behbahani HF, Takeno T. An excess enthalpy flame combustor forextended flow ranges. In: Proceedings of the twentieth symposium(international) on combustion. The Combustion Institute; 1984. p. 2025–33.
[48] Barra AJ, Ellzey JL. Heat recirculation and heat transfer in porous burners.Combust Flame 2004;137:230–41.
[49] Min DK, Shin HD. Laminar premixed flame stabilized inside a honeycombceramic. Int J Heat Mass Trans 1991;34:341–56.
[50] Hanamura K, Echigo R. An analysis of flame stabilization mechanisms inradiation burners. Warme Stoffubertrag 1991;26:377–83.
[51] Hsu P-F, Evans WD, Howell JR. Experimental and numerical study ofpremixed combustion within nonhomogeneous porous ceramics. CombustSci Technol 1993;90:149–72.
[52] Hsu P-F, Howell JR, Matthews RD. A numerical investigation of premixedcombustion within porous inert media. J Heat Transf 1993;115:744–50.
[53] Khanna V, Goel R, Ellzey JL. Measurement of emissions and radiation formethane combustion within a porous medium burner. Combust Sci Technol1994;99:133–42.
[54] Kulkarni MR, Peck RE. Analysis of a bilayered porous radiant burner. NumerHeat Trans A—Appl 1996;30:219–32.
[55] Mathis Jr WM, Ellzey JL. Flame stabilization, operating range, and emissionsfor a methane/air porous burner. Combust Sci Technol 2003;175:825–39.
[56] Vogel BJ, Ellzey JL. Subadiabatic and superadiabatic performance of a two-section porous burner. Combust Sci Technol 2005;177:1323–38.
[57] Golombok M, Prothero A, Shirvill LC. Surface combustion in metal fibreburner. Combust Sci Technol 1991;77:203–23.
[58] Itaya Y, Miyoshi K, Maeda S, Hasatani M. Surface combustion of a premixedmethane–air gas on a porous ceramic. Int Chem Eng 1992;32:123–31.
[59] Kawaguchi O, Otoh T, Nakamura S, Todoroki A, Murayama Y. Premixedcombustion at a fiber mat. In: Proceedings of the twentieth symposium(international) on combustion. The Combustion Institute; 1990. p. 1019–24.
[60] Bouma PH, De Goey LPH. The stability of flames on a ceramic foam surfaceburner. In: Proceedings of the first European conference on small burnertechnology and heating equipment; 1996. p. 309–18.
[61] Bouma PH, De Goey LPH. Premixed combustion on ceramic foam burners.Combust Flame 1999;119:133–43.
[62] Bouma PH, Eggels RLGM, De Goey LPH, Nieuwenhuizen JK, Van der Drift A.A numerical and experimental study of the NO-emission of ceramic foamsurface burners. Combust Sci Technol 1995;108:193–203.
[63] Bouma PH, Eggels RLGM, De Goey LPH. NO emissions of ceramic foamsurface gas burners in the radiation mode. In: Proceedings of the thirdinternational conference on combustion technologies for a clean environ-ment (Clean Air III). Lisbon; 1995.
[64] Bouma PH, Somers LMT, De Goey LPH, Nieuwenhuizen JK. Modelling ofmethane-air combustion on ceramic foam surface burners in the radiationmode. Heat Technol 1995;13:127–39.
[65] Schreel KRAM, van den Tillaart EL, de Goey LPH. The influence of burnermaterial properties on the acoustical transfer function of radiant surfaceburners. In: Proceedings of the thirtieth symposium (international) oncombustion. The Combustion Institute; 2005. p. 1741–8.
[66] Buckmaster J, Takeno T. Blow-off and flashback of an excess enthalpy flame.Combust Sci Technol 1981;25:153–8.
[67] Afsharvahid S, Dally BB, Christo FC. Numerical study of flame stabilizationinside porous burner. In: Proceedings of the 2003 Australian symposium oncombustion; 2003.
[68] Sathe SB, Peck RE, Tong TW. A numerical analysis of heat transfer andcombustion in porous radiant burners. Int J Heat Mass Trans 1990;33:1331–8.
[69] Afsharvahid S, Dally BB, Christo FC. On the stabilisation of ultra-leanmethane and propane flames in porous media. In: Proceedings of the fourthAsia-Pacific conference on combustion; 2003.
[70] Mital R, Gore JP, Viskanta R. A study of the structure of submergedreaction zone in porous ceramic radiant burners. Combust Flame 1997;111:175–84.
[71] Sathe SB, Kulkarni MR, Peck RE, Tong TW. An experimental and theoreticalstudy of porous radiant burner performance. In: Proceedings of the twenty-third symposium (international) on combustion. The Combustion Institute;1990. p. 1011–8.
[72] Smucker MT, Ellzey JL. Computational and experimental study of a two-section porous burner. Combust Sci Technol 2004;176:1171–89.
[73] Malico I, Pereira JCF. Numerical predictions of porous burners withintegrated heat exchanger for household applications. J Porous Media 1999;2:153–62.
[74] Henneke MR, Ellzey JL. A numerical study of re-ignition in porous mediaburners. ASME HTD 1997;341:155–63.
[75] Aldushin AP. New results in the theory of filtration combustion. CombustFlame 1993;94:308–20.
[77] Babkin VS, Korzhavin AA, Bunev VA. Propagation of premixed gaseousexplosion flames in porous media. Combust Flame 1991;87:182–90.
[78] Bingue JP, Saveliev AV, Fridman AA, Kennedy LA. NOX and CO emissions oflean and ultra-lean filtration combustion of methane/air mixtures in an inertporous media. Clean Air 2002;3:199–210.
[79] Contarin F, Saveliev AV, Fridman AA, Kennedy LA. A reciprocal flow filtrationcombustor with embedded heat exchangers: numerical study. Int J HeatMass Trans 2003;46:949–61.
ARTICLE IN PRESS
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684 683
[80] Di Mare L, Mihalik TA, Continillo G, Lee JHS. Experimental and numericalstudy of flammability limits of gaseous mixtures in porous media. ExpTherm Fluid Sci 2000;21:117–23.
[81] Dobrego KV, Kozlov IM, Bubnovich VI. Dynamics of filtration combustionfront perturbation in the tubular porous media burner. Int J Heat Mass Trans2003;46:3279–89.
[82] Dobrego KV, Gnesdilov NN, Kozlov IM, Bubnovich VI, Gonzalez HA.Numerical investigation of the new regenerator-recuperator scheme ofVOC oxidizer. Int J Heat Mass Trans 2005;48:4695–703.
[83] Fouteko SI, Shabunya SI, Zhdanok SA, Kennedy LA. Superadiabatic combus-tion wave in a diluted methane–air mixture under filtration in a packed bed.In: Proceedings of the twenty-sixth symposium (international) on combus-tion. The Combustion Institute; 1996. p. 3377–82.
[84] Hanamura K, Echigo R, Zhdanok SA. Superadiabatic combustion in a porousmedium. Int J Heat Mass Trans 1993;36:3201–9.
[85] Henneke MR, Ellzey JL. Modeling of filtration combustion in a packed bed.Combust Flame 1999;117:832–40.
[86] Hoffmann JG, Echigo R, Yoshida H, Tada S. Experimental study oncombustion in porous media with a reciprocating flow system. CombustFlame 1997;111:32–46.
[87] Jugjai S, Chuenchit C. A study of energy conversion by a porous combustor-heat exchanger with cyclic flow reversal combustion. Int Energy J 2001;2:95–104.
[88] Kakutkina NA, Korzhavin AA, Mbarawa M. Filtration combustion ofhydrogen–air, propane–air and methane–air mixtures in inert porous media.Combust Explos Shock 2006;42:372–83.
[89] Kennedy LA, Fridman AA, Saveliev AV. Superadiabatic combustion in porousmedia: wave propagation, instabilities, new type of chemical reactor. FluidMech Res 1995;22:1–26.
[90] Kennedy LA, Bingue JP, Saveliev AV, Fridman AA, Foutko SI. Chemicalstructures of methane-air filtration combustion waves for fuel-lean and fuel-rich conditions. In: Proceedings of the twenty-eighth symposium (interna-tional) on combustion. The Combustion Institute; 2000. p. 1431–8.
[91] Kennedy LA, Saveliev AV, Bingue JP, Fridman AA. Filtration combustion of amethane wave in air for oxygen-enriched and oxygen-depleted environ-ments. In: Proceedings of the twenty-ninth symposium (international) oncombustion. The Combustion Institute; 2002. p. 835–41.
[92] Korzhavin AA, Bunev VA, Abdullin RK, Babkin VS. Flame zone in gascombustion in an inert porous medium. Combust Explos Shock 1982;18:628–31.
[93] Zhang G, Cai X, Liu M, Lin B, Chen Y, Wang L. Characteristic analysis of low-velocity gas filtration combustion in an inert packed bed. Combust TheorModel 2006;10:683–700.
[94] Zhdanok S, Kennedy LA, Koester G. Superadiabatic combustion of methaneair mixtures under filtration in a packed bed. Combust Flame1995;100:221–31.
[95] Babkin VS. Filtrational combustion of gases: present state of affairs andprospects. Pure Appl Chem 1993;65:335–44.
[96] Babkin VS, Laevskii YM. Seepage gas combustion. Combust Explos Shock1987;23:531–47.
[97] Rumminger MD, Dibble RW, Heberle NH, Crosley DR. Gas temperature abovea porous radiant burner: comparison of measurements and model predic-tions. In: Proceedings of the twenty-sixth symposium (international) oncombustion. The Combustion Institute; 1996. p. 1755–62.
[98] Miller JA, Bowman CT. Mechanism and modeling of nitrogen chemistry incombustion. Prog Energy Combust 1989;15:287–338.
[99] Lim I-G, Matthews RD. Predictions of NOX and CO emissions from porousmedia burners. ASME PD 1993;50:77–84.
[100] Delalic N, Mulahasanovic D, Ganic EN. Porous media compact heatexchanger unit—experiment and analysis. Exp Therm Fluid Sci 2004;28:185–92.
[101] Mital R, Gore JP, Viskanta R. A radiation efficiency measurement procedurefor gas-fired radiant burners. Exp Heat Transfer 1998;11:3–21.
[102] Chaffin C, Koenig M, Koeroghlian M, Matthews RD, Hall MJ, Nichols SP, et al.Experimental investigation of premixed combustion within highly porousmedia, In: Proceedings of the ASME/JSME thermal engineering jointconference, Reno, 1991. p. 219–24.
[103] Ellzey JL, Goel R. Emissions of CO and NO from a two stage porous mediaburner. Combust Sci Technol 1995;107:81–91.
[104] Afsharvahid S, Lanspeary PV, Dally BB, Christo FC. On the design of an ultralean porous burner. In: Proceedings of the 2002 Australian symposium oncombustion; 2002.
[106] Fend T, Trimis D, Pitz-Paal R, Hoffschmidt B, Reutter O. Thermal properties.In: Scheffler M, Colombo P, editors. Cellular ceramics: structure, manufac-turing, properties and applications. Weinheim: Wiley; 2005. p. 342–60.
[107] Barra AJ, Diepvens G, Ellzey JL, Henneke MR. Numerical study of the effectsof material properties on flame stabilization in a porous burner. CombustFlame 2003;134:369–79.
[108] Fu X, Viskanta R, Gore JP. Modeling of thermal performance of a porousradiant burner. ASME HTD 1998;361:11–8.
[109] Pickenacker O, Pickenacker K, Wawrzinek K, Trimis D, Pritzkow WEC, MullerC, et al. Innovative ceramic materials for porous-medium burners.Interceram 1999;48:326–30.
[110] Pickenacker O, Pickenacker K, Wawrzinek K, Trimis D, Pritzkow WEC, MullerC, et al. Innovative ceramic materials for porous-medium burners II.Interceram 1999;48:424–33.
[111] Bowen HK. Basic research needs on high temperature ceramics for energyapplications. Mater Sci Eng 1980;44:1–56.
[112] Xiong T-Y, Khinkis MJ, Fish FF. Experimental study of a high-efficiency, lowemission porous matrix combustor-heater. Fuel 1995;74:1641–7.
[113] Brenner G, Pickenacker K, Pickenacker O, Trimis D, Wawrzinek K, Weber T.Numerical and experimental investigation of matrix-stabilized methane/aircombustion in porous inert media. Combust Flame 2000;123:201–13.
[114] Kamal MM, Mohamad AA. Development of a cylindrical porous-mediumburner. J Porous Media 2006;9:469–81.
[115] Pickenacker O, Trimis D. Experimental study of a staged methane/air burnerbased on combustion in a porous inert medium. J Porous Media 2001;4:197–213.
[117] Kamal MM, Mohamad AA. Enhanced radiation output from foam burnersoperating with a nonpremixed flame. Combust Flame 2005;140:233–48.
[118] Kesting A, Pickenacker K, Trimis D, Cerri I, Krieger R, Schneider H.Development of a highly efficient gas infrared heater by means ofcombustion in inert porous media. In: Proceedings of the international gasresearch conference; 2001.
[119] Schmidt J, Scheiffele M, Mach A, von Issendorff F. Highly porous non-oxideceramic matrix composites (C/SiSiC and SiSiC) for porous burner applica-tions. In: Proceedings of the 2005 conference on porous ceramic materials(PCM2005); 2005.
[120] Cannon RW, Gugel E, Leimer G, Woetting G. Ceramics, advanced structuralproducts. In: Ullmann F, editor. Ullmann’s enclyclopedia of industrialchemistry. 7th ed. New York: Wiley; 2000.
[121] DePoorter GL, Brog TK, Readey MJ. Properties and applications of structuralceramics. In: Properties and selection: nonferrous alloys and special-purpose materials. ASM handbook, vol. 2. Ohio: ASM International; 1992.
[122] Meng WH, McCordic C, Gore JP, Herold KE. A study of effects of porousceramic inserts on heat transfer and combustion in a fired heat exchanger.In: Proceedings of the third ASME/JSME thermal engineering jointconference. Reno; 1991. p. 181–8.
[123] Leonardi SA, Gore JP, Viskanta R. Experimental investigation of partially-premixed combustion in a novel porous radiant burners. In: Proceedings ofthe thirty-fifth national heat transfer conference (NHTC2001). Anaheim;2001. p. 915–21.
[124] Huang Y, Chao CYH, Cheng P. Effects of preheating and operation condi-tions on combustion in a porous medium. Int J Heat Mass Trans 2002;45:4315–24.
[125] Porvair. Metal foam. Available from: /http://www.porvairadvancedmaterials.comS. Last accessed: 18 March 2008.
[126] Lei GY. Basic metallurgy and product forms. In: Heat treating, vol. 2. Ohio:ASM Handbook. ASM International; 1992.
[127] MatWeb. Fecralloy TM 145 electrical resistance steel. Available from:/http://www.matweb.comS. Last accessed: 18 March 2008.
[128] Marbach TL, Agrawal AK. Experimental study of surface and interiorcombustion using composite porous inert media. J Eng Gas Turb Power2005;127:307–13.
[129] Gibson LJ, Ashby MF. Cellular solids: structure and properties. 2nd ed.Cambridge: Cambridge University Press; 1997.
[130] Elverum PJ, Ellzey JL, Kovar D. Durability of YZA ceramic foams in a porousburner. J Mater Sci 2005;40:155–64.
[131] Schmidt J, Scheiffele M, Mach A, von Issendorff F, Ortona A. Manufacturing,characterization and testing of SiSiC foams for porous burner applications.In: Proceedings of the 2005 conference on porous ceramic materials(PCM2005); 2005.
[132] Mach A, Herzog A, von Issendorff F, Kanka B, Krieger R, Moßbauer S, et al.CERPOR project: optimization of ceramic components for the porous burnertechnology. In: Proceedings of the 2005 conference on porous ceramicmaterials (PCM2005); 2005.
[133] San-Gobain NorPro. Heat sink and heat transfer media products andmaterials. Available from: /http://www.norpro.saint-gobain.comS. Lastaccessed: 18 March 2008.
[134] Durst F, Trimis D. Brenner. Ger Offen DE 4322109.2 1995.[135] Hsu P-F. Experimental study of the premixed combustion within the
nonhomogeneous porous ceramic media. ASME HTD 1996;329:1–10.[136] Mohamad AA. Numerical simulation of combustion in a cylindrical porous
medium. In: Proceedings of the 2002 engineering technology conference onenergy (ETCE 2002); 2002. p. 85–91.
[137] Mohamad AA. Fundamentals of combustion in porous burners. In: Proceed-ings of the NATO advanced study institute on porous media. Neptun-Olimp;2003. p. 348–57.
[138] Zhdanok SA, Dobrego KV, Futko SI. Flame localization inside axis-symmetriccylindrical and spherical porous media burners. Int J Heat Mass Trans1998;41:3647–55.
[139] Brockerhoff P, Emonts B. Test of a premixing radiant burner for the low NOX
combustion of natural gas/hydrogen mixtures. Int J Hydrogen Energy1994;19:395–8.
[140] Kamal MM, Mohamad AA. Effect of swirl on performance of foam porousmedium burners. Combust Sci Technol 2006;178:729–61.
S. Wood, A.T. Harris / Progress in Energy and Combustion Science 34 (2008) 667–684684
[141] Viskanta R. Enhancement of heat transfer in industrial combustion systems:problems and future challenges. In: Proceedings of the third ASME/JSMEthermal engineering joint conference, vol. 5. Reno; 1991. p. 161–73.
[142] Echigo R, Kurusu M, Ichimiya K, Yoshizawa Y. Combustion augmentation ofextremely low calorific gases (application of the effective energy conversionmethod from gas enthalpy to thermal radiation). In: Proceedings of theASME/JSME thermal engineering joint conference, vol. 4. Honolulu; 1983. p.99–103.
[143] Jugjai S, Sawananon A. The surface combustor-heater with cyclic flowreversal combustion embedded with water tube bank. Fuel 2004;83:2369–79.
[144] Durst F, Weclas M. A new type of internal combustion engine based on theporous-medium combustion technique. P I Mech Eng D—J Aut 2001;215:63–81.
[145] Hanamura K, Bohda K, Miyairi Y. A study of super-adiabatic combustionengine. Energy Convers Manage 1997;38:1259–66.
[146] Hanamura K, Kumano T, Iida Y. Electric power generation by super-adiabaticcombustion in thermoelectric porous element. Energy 2005;30:347–57.
[147] Catalano A. Thermophotovoltaics: a new paradigm for power generation?Renew Energy 1996;8:495–9.
[148] Qiu K, Hayden ACS. Thermophotovoltaic power generation systems usingnatural gas-fired radiant burners. Sol Energy Mat Sol C 2007;91:588–96.
[149] Duran A, Rost M. Flameless thermal oxidation for chlorinated waste vent gasdestruction: a case history. In: Proceedings of the twenty-fourth interna-tional conference on incineration and thermal treatment technologies (IT32005). Galveston; 2005.
[150] Martin RJ, Stilger JD, Holst MR. Method and apparatus for controlled reactionin a reaction matrix. US Patent 5,165,884; 24 November 1992.
[151] Noordally E, Przybylski JM, Witton JJ. Porous media combustors for clean gasturbine engines. Report 9210/1-01/04. Bedford: Cranfield University; 2004.
[152] Altendorfner M, von Issendorf F, Dannehl M. Development of a high pressureburner for industrial applications using inert porous media. In: Proceedingsof the ninth conference on energy for a clean environment (Clean Air IX).Povoa de Varzin; 2007.
[153] Moßbauer S, Gruber W, Trimis D. Exhaust gas recirculation in porousburners for the target application zero emission steam engines. In:Proceedings of the sixth international conference on technologies combus-tion for a clean environment (Clean Air VI); 2001.
[154] Bell RD, Chaffin C, Koeroghlian M. Experimental investigation of a stagedporous ceramic burner. ASME PD 1992;39:41–6.
