ASHRAE ST-16-012-2016 Optimizing Excess Air in Relation to Energy Temperature and Reduction of Emissions of Methane Gas in a Combustion Nozzle- Using Numerical Combustion Modeling.pdf

1、 2016 ASHRAE 115ABSTRACTOptimizing the combustion performance and reducingemissions of methane gas by varying excess air has been andcontinues to be an area of interest for researchers, manufac-turers, and operators of equipment such as gas-fired boilers,with the aim of developing more energy-effici

2、ent systems andaddressing ever-growing environmental concerns. Resultswere graphically displayed and discussed, illustrating theeffects of excess air on the combustion process with regards toenergy, temperature, and pollutants. Easy-to-use equationswere developed with guidance on how to accurately o

3、ptimizecombustion.Methodology: Numerical software tools were used inanalyzing injected methane gas and variable excess air ratios.Emissions such as carbon dioxide (CO2), carbon monoxide(CO), and nitrogen oxides (NOx) were also recorded andanalyzed. Optimum energy output was investigated in relationt

4、o excess air and emissions.Results were tabulated and graphs generated. Equationswere derived using industry-established software tools. Theaccuracy of the developed equations was assessed on statisti-cal basis. Discussions on advantages and disadvantaged onexcess air are included.Conclusion: Result

5、s were found to be in agreement withpublished official information. Precise control of excess aircanimproveenergyefficiencyandlowerpollutionlevelsintheflue gas. The developed equations can easily be programmedinto a computer-controlled combustion system. Results in thispaper can be also be used to a

6、ssess similar existing processes.Oncethecomputationalfluiddynamics(CFD)modelissetup,the model can be adjusted to suit specific simulation require-ments.INTRODUCTIONStudying the effects of excess air on combustion effi-ciency in relation to polluting emissions is becoming evermore important. This int

7、erest is being driven by depletinghydrocarbon fuels, rising fuel costs, and environmentalconcerns. Environmental bodies, manufacturers, and endusers of gas-fired equipment are expected to become moreinterested in this technology.Combustornozzlesofdifferentcharacteristicsareusedingas turbines and wat

8、er heating boilerstwo examples oper-atingunderdifferentparameters.Optimizingperformanceandreducing emissions of such combustion processes can posi-tively impact the two mentioned examples.The numerical research aims to demonstrate howcombustionenergyoutputbehavesoverawiderangeofair-to-fuel ratios. I

9、mpact of variable air-to-fuel ratios on nitrogenoxides (NOx), carbon monoxide (CO), and carbon dioxide(CO2) was included in this paper with graphical contourpictures and graph plots.Curve-fitting techniques were employed to mathemati-cally describe the relationship between injected methane gasand va

10、rying injected airflow rates. The developed equationlendsitselftocomputerprogramming.Acomputer-controlledcombustionprocesscanaccuratelybecontrolledinoptimizingtotal energy output and total temperature output, and thereduction of CO, CO2, and NOx. Equations developed byusing curve-fitting techniques

11、were statically assessed forintegrity.Optimizing Excess Air in Relation to Energy,Temperature,andReductionofEmissionsofMethane Gas in a Combustion NozzleUsing Numerical Combustion ModelingAli M. Hasan, CEngMember ASHRAEAli M. Hasan is a senior engineer at KEO International Consulting Engineers, Doha

12、, Qatar.ST-16-012Published in ASHRAE Transactions, Volume 122, Part 2 116 ASHRAE TransactionsEFFECTS OF VARYING INJECTED AIRFLOW ONMETHANE GAS IN A COMBUSTION NOZZLEIngeneral,flametemperatureishighestwhenthereisjustenough fuel to react with the available oxygen. This process isknown as stoichiometri

13、c combustion, producing a fast chem-ical reaction with a high flame temperature. Nonstoichiomet-ric combustion is a process with excess air (lean mixture) andexcess fuel (rich mixture). A rich mixture produces lowertemperatureandreactionrates.However,inpractice,completeorstoichiometriccombustioncanb

14、edifficulttoachieve,caus-ingharmfulunburnedfuelgasesandCOgasestoescapetotheatmosphere.Excess air is therefore important but within limits,becausehighexcessairusedinanendeavortoobtaincompletecombustion absorbs some of the combustion heat, causing adrop in efficiency. The aim will be to find the optim

15、um excessair value.European and American definitions (Kutz 2006): TheAmerican definition refers to equivalence ratio, whichdescribes the ratio of fuel to air relative to the stoichiometriccondition. An equivalence ratio of 1.0 corresponds to the stoi-chiometric condition. At lean fuel conditions the

16、 ratio is lessthan 1. Conversely, at fuel-rich conditions it is greater than 1.The European definition refers to the reciprocal, which is thelambda value .The combustion model was assumed to be a nonadiabaticprocess. Process energy and temperature output will alsodepend on the following parameters d

17、escribed in the “Meth-odology” section.METHEDOLOGYThe following settings and assumptions were made usingANSYS Fluent software (ANSYS 2016):Non-adiabatic process.Air-to-fuel ratio: 17.24 kg to 1 kg of fuel, (37.93 lb to2.2 lb of fuel). Ratio for stoichiometric conditionsobtained from ASHRAE HandbookF

18、undamentals(ASHRAE 2009). The model was repeated with differ-ent air-to-fuel ratios, as shown in Table 1.Fluent standard database air fractions: 0.78992 of nitro-gen and 0.21008 of oxygen.K-epsilon model, a common turbulence model. Accord-ing to Spalding (2014), the k-epsilon turbulence modelis a go

19、od compromise between generality and economyof use for many computational fluid dynamics (CFD)problems.P1 radiation model. According to Liu et al. (2009), theP1 model is the simplest case of the more general P-Nmodel. For combustion and complicated geometries, theP1 model works reasonably well. A pa

20、per by Prieler etal. (2014), which simulates an oxygen-natural gas com-bustion process, states that the P1 model overestimatesradiation, which leads to a lower temperature in the fur-nace.Non-premixed fuels.Air and fuel temperatures at inlet: 300 K. Table 1 showsapplied mass flow rates.Automatic mes

21、hing tool was selected.Number of iterations made: 1500. At this setting, theprocessor was monitored and the residuals plot showedstability. Residual curves flattened out halfway throughprocessing.Probability density function (PDF) was used to simulatecombustion. According to Pope (1997), the PDF mod

22、elcan produce satisfactory results for turbulent reactionflows where convection effects due to mean and fluctuat-ing components of velocity are dominant. The model canbe extended for adiabatic and non-adiabatic conditions.The Zeldovich mechanism was used to calculate NOxformation. See the “NOxby Mas

23、s Fraction” section.Mesh metricelement quality: 388,085 elements wereused. Minimum Jacobian element was 0.40, maximumJacobian element was 0.99, average Jacobian element wasTable 1. Varying Inlet Air Flow and Combustion Output of Methane GasInlet AirflowCombustionTotal Temperature Total Energy NOxCO

24、CO2Excess Air, % e-3, kg/s (lb/s) e2, K e4, J/kg (Btu/lb) e-5, Mass Fraction e-2, Mass Fraction e-2, Mass Fraction0 17.24 (37.93) 21.37 15.0408 (64.6638) 13.8 14.05 18.10 (39.82) 21.43 17.0738 (73.4041) 29.21 13.8 13.810 18.96 (41.71) 21.37 17.0334 (73.2304) 26.12 13.9 13.820 20.69 (45.52) 21.36 15.

25、2554 (65.5864) 18.01 13.8 13.730 22.41 (49.30) 21.23 12.5216 (53.8332) 10.69 13.6 14.040 24.14 (53.11) 21.01 9.3204 (40.0705) 5.062 13.4 13.960 27.58 (60.67) 20.46 3.1713 (13.6341) 2.110 0.132 13.9Note: Results were obtained with 1 e-3 kg/s (2.2 e3 lb/s) of methane gas injected at the combustor inle

26、t. = 1 refers to 0% excess air and = 1.6 refers to 60% excess air.1 kJ/kg = 0.4299 Btu/ lb. 1 kg/s = 2.2lb/s.Published in ASHRAE Transactions, Volume 122, Part 2 ASHRAE Transactions 1170.90, and the standard deviation of Jacobian element was7.02 e-2. An Jacobian element close to 1 indicates a goodel

27、ement quality, while an Jacobian element close to 0 indi-cates a poor element quality.Items not mentioned above were left at software defaultsettings. A model of the injectors can be seen in Figure 1, withmodel description. CFD simulations are parametric and caneasilybeadaptedtosimulateothergeometry

28、orflowrates.Foranexample, to change the diameter of the injectors shown inFigure 1,adjustthediameteratthesoftwaremodeler,thenremesh(checkqualityofmeshasinthelistabove),thenmovetothesetupstage/update data, and then re-run the processor. Follow proce-dure as discussed for curve fittings.RESULTS AND DI

29、SCUSSIONSThe software database was used to select and define theproperties of methane gas fuel and air. Figures 2 through 6 areassociated with total temperature, total energy, NOx, CO, andCO2,respectively.Eachofthereferred-tofiguresshowtheimpactof excess air on the combustion process. Contour plots

30、provide asnapshot within the combustion process with regards to theselected property output.Associated software input data and the output results areshown in Table 1. Figure 7 shows the plotted results obtainedfrom Table 1. MicrosoftExcel2010 software tools wereutilizedtoplotFigure8andusecurve-fitti

31、ngtoolstogeneratethedeveloped equation.Note:TotalenergywasmeasuredinJ/kg(Btu/lb)ofcombus-tion output. NOx, CO, and CO2were measured as fraction ofmass in the nozzle combustion output. R2describes the integrityof formulas developed by regression. With a low R2= 0, none ofthevariancesontheyaxiscanbeex

32、plainedagainstpercentexcessair (EA) shown on the x axis. With a high R2= 1, all of the vari-ances on the y axis can be explained against percent EA on the xaxis.(b)Figure 1 (a) Disc showing methane gas and air injectors.The disc outer diameter is 100 mm (4 in.) and is10 mm (0.04 in.) thick. The disc

33、 is metal with aceramic heat-resisting coating. (b) Combustionfluid geometry considered showing side and frontviews. DN represents the injector diameters (DN5is 5 mm 0.2 in.).(a)(b)Figure 2 Combustiontotaltemperature.(a) =1.05excessair. (b) =1.60 excess air. Maximum values canbe seen in Table 1.(a)P

34、ublished in ASHRAE Transactions, Volume 122, Part 2 118 ASHRAE TransactionsTotal EnergyTotal energy increased as the temperature of air inletincreased, as shown in Figures 2 and 7. Curve fitting wasemployed using Excel 2010.The curve in Figure 7 shows the total energy output atstoichiometric conditi

35、ons or = 1. Total energy peaks at 5%excess air and begins to drop. At 20% excess air, total energyis similar to excess air = 1 condition. It is recommended thatthe following equation be used within the range of percentageexcess air as shown in Table 1: 0% to 60%.Total energy = 0.0057(EA)4+ 0.5345(EA

36、)318.749(EA)2+ 289.51(EA) 1639.4 (1)with a regression R2= 0.9991. This represents a good level ofapproximation.It is recommended that Equation 1 be used with thefollowing condition:01.2or with equivalence ratios 0 0.83 (2)Total TemperatureFigures 3 and 7 show that as excess air increases, thecombust

37、ion temperature decrease. This is expected andexplainedinthe“EffectsofVaryingInjectedAirflowonMeth-ane Gas in a Combustion Nozzle” section. It is recommendedthat the following equation be used within the range ofpercentage excess air as shown in Table 1: 0% to 60%.Total temperature = 0.0113(EA)2+ 0.

38、4156(EA) + 17.572 (3)with a regression R2= 0.9961.Table 1 and Figure 7 show that the total temperaturebegins to drop at = 1.2 and above.(b)Figure 3 Combustion total energy. (a) = 1.05 excess air.(b) = 1.60 excess air. Maximum values can beseen in Table 1.(a)(a)(b)Figure 4 NOxcombustion mass fraction

39、. (a) = 1.05excess air. (b) = 1.60 excess air. Maximumvalues can be seen in Table 1.Published in ASHRAE Transactions, Volume 122, Part 2 ASHRAE Transactions 119CO by Mass FractionTable 1 and Figures 4 and 7 show that CO levels decreaseas excess air increases. However, it is important to refer to the

40、conditions mentioned in the Equation 2. The designer or oper-ator can refer to the abovementioned equation to select whatis best suitable and comply with any local government restric-tionsonemissions.Itisrecommendedthatthefollowingequa-tionbeusedwithintherangeofpercentageexcessairasshownin Table 1:

41、0% to 60%.CO = 0.0043(EA)2+ 0.1264(EA) + 12.959 (4)with a regression R2= 0.9249.For maximum efficiency, combustion at low excess air isdesirable (ASHRAE 2009). Figure 7 indicates a minimal dropin CO levels between = 1 and = 1.2. More significant COdrop was observed at = 1.2 and above. The designer a

42、ndoperatorcanrefertothedevelopedequationsandadjustexcessair to achieve any specified CO levels, knowing how totalenergy and total temperature levels are impacted.CO2by Mass FractionTable 1 and Figure 7 show that CO2mass fraction hasremained relatively unchanged as the air-to-fuel ratioincreased. Lev

43、els of CO2are expected to remain constant onthe basis of carbon content in fuel, but when considering CO2in relation to the increased excess air in the exhaust, then CO2fractionofexhaustflueisexpectedtodrop.Itisrecommendedthattheequationinthefollowingparagraphbeusedwithintherangeofpercentageexcessai

44、rasshowninTable1:0%to60%.Note: Comparing the mass fraction of CO2= 0.148 withpublished figures, 0.148 100 (converting to a percentage)(a)(b)Figure 5 CO combustion. (a) = 1.05 excess air.(b) = 1.60 excess air. Maximum values can beseen in Table 1.(b)Figure 6 CO2combustion. (a) = 1.05 excess air. (b)=

45、 1.60 excess air. Maximum values can beseen in Table 1.(a)Published in ASHRAE Transactions, Volume 122, Part 2 120 ASHRAE Transactionsgives 14.8%. This number is close to what is indicated inASHRAE HandbookFundamentals (ASHRAE 2009),which is 12% at theoretical stoichiometry. This comparisoncan also

46、be regarded as a check on level of accuracy betweenresults obtained and what is published.NOxby Mass FractionTable 1 and Figure 7 show that levels of NOxincrease asexcess air increases together with peak total energy output,then drop as the total energy and total temperature outputbegin to drop. The

47、 peak increases can be explained by theexcessive availability of nitrogen in the increased excess air.The drop can be explained by the beginning of temperaturedrop. According to Kutz (2006), thermal NOxis generallyregarded as being generated by a chemical reaction sequencecalled the Zeldovich mechan

48、ism, and the rate of NOxforma-tion is proportional to temperature.NOxmass fraction = 0.0407(EA)4+ 3.7382(EA)3127.58(EA)2+ 1913.5(EA) 10612 (5)with a regression R2= 0.9817.In this equation and as shown in Figure 8, the NOxcurvefittingshowsdiversionfromtheactualcurveathighexcessairlevels, at = 1.6 and above. Limiting the production of NOxcan be done with a precise control of excess air in a combus-tion process. The condition developed in