Prediction of NO X Emissions

Size: px

Start display at page:

Download "Prediction of NO X Emissions"

Angelica Wilkins
5 years ago
Views:

1 p ;A$ 7-2: 81-GT-119 TH AMRICAN SCITY F MCHANICAL NGINRS St., New York, N.Y. 117 The Society shll not be responsible for sttements or opinions dvnced in ppers of in discussion t meetings of the Society or of its Divisions or Sections, or printed in its publictions. Discussion is printed only if the pper is published in n ASM Journl or Proceedings. Relesed for generl publiction upon presenttion,fuli credit should be given to ASM. the Techmct Division. nd the uthorisl. Copyright 1981 by ASM G. D. Lewis Prtt & Whitney Aircrft, Government Products Division, West Plm Bech, Fl. Prediction of N X missions A procedure hs been developed for predicting the formtion of N, in gs turbine combustion systems operting t vrious conditions nd on vriety of fuels, if the N,r produced by tht engine is known for ny one combintion of fuel nd operting condition. The predictions re bsed on fundmentl reltionship between N, formtion nd flme temperture with empiricl djustments for the specil cses of fuel chnges nd wter injection. ngine dt indicte tht the predictions re ccurte within the dt sctter limits tht re normlly encountered in engine field mesurements. This work is not intended s new kinetic model of N, formtion but rther s useful tool for the engineer who hs to predict wht his engine will do under conditions where it hs not yet been tested. It divides the N, formtion process into fctors mesurble by the engineer, i.e., flme temperture, pressure, nd fuel flow rte, nd shows how ech of these contributes to the overll N, in predictble wy. DISCUSSIN N formtion is well-known to be strong function of temperture. A reltionship between combustion tempertures nd N. concentrtion ws developed from dt vilble in the literture nd tested for vlidity ginst engine dt. The effects of fuel type, wter injection, stem injection nd power level on N,r genertion were found to be predictble when dditionl llownces were mde for the effects of pressure nd fuel flow rte chnges. In prticulr, it ws found tht the lower N x emissions experienced with nturl gs fuel s compred to light petroleum distillte is due to the lower pek flme temperture of nturl gs nd is not cused by its vpor stte t injection. Although the quntity of vilble dt for one engine operting on two different fuels or with nd without diluent injection ws smll, the correltion between the predicted nd mesured vlues is good. FUNDAMNTAL RLATINSHIP The excellent mesurements of N x formtion in n homogenous flme mde by Anderson t the NASA Lewis Reserch Center [1] formed the bsis for the fundmentl reltionship presented in Fig. 1. The experimentl work ws conducted in premixed combustion system using propne fuel nd ir t 5.5 tmospheres pressure nd two different inlet tempertures. Although Anderson presented the dt s plot of N x emission index vs equivlence rtio t two inlet tempertures, they hve been converted here to prts per million by volume (PPMV) vs flme temperture clculted from the equivlence rtios nd inlet ir tempertures. In ddition, the originl N x mesurements t 5.5 tmospheres hve been djusted to more conventionl gs turbine burner pressure of 2 tmospheres by the generlly ccepted squre root of pressure reltionship. The dt on the figure represent experimentl mesurements t equivlence rtios below.8. At higher equivlence rtios, the deficiency of oxygen cused the dt to devite from the stright line shown. This line, therefore, represents the N 5 formed immeditely s the fuel frgments re burned nd is function of temperture only. Although the dt were obtined from len, fuel vpor-ir mixture, they represent the effect of temperture on N x formtion nd cn be used to predict the chnge in N 5 formtion in given combustion system, rich or len, if only the temperture is chnged, such s by chnging the inlet ir temperture. These N x mesurements were obtined fter flme residence time of 2 msec, but both the dt of Anderson [1] nd Hrris, et l., [2] indicte tht the N,, forms immeditely s the combustion occurs nd tht resonbly longer residence times do not gretly ffect the results. The line through the dt in Fig. 1 hs been extrpolted (dshed line) nd cn be represented by the eqution: N x = 7.5 X 1-s e(8.28xs-)(t) (1) where T is the flme temperture in K. For len, homogenous, prevporized combustion systems, this eqution cn be used directly to predict N 5 s ws demonstrted []. For equivlence rtios bove.8 or for nonhomogenous flmes (droplet burning or diffusion gseous-fuel flmes), however, oxygen depletion prohibits the use of eqution 1 to predict N x directly. In both the gseous fuel diffusion flme nd droplet burning, the combustion tkes plce very ner stoichiometric fuel-ir rtio. ven t these conditions, however, eqution 1 cn be used to predict the effect of chnges in flme temperture on N. formtion, even though the bsolute level of N,, hs been reduced by the oxygen deficiency. The rtio of Nx produced t temperture T 1 to tht produced t temperture T 2 cn be clculted by dividing eqution 1 contining T i by eqution 1 contining T 2. This reduces to the form of eqution 2: N x2 = N Rtio = 8(8.28x1-)(T2-Tl). ( 2) N, Contributed by the Gs Turbine Division of TH AMRICAN SCITY F MCHANICAL NGINRS for presenttion t the Gs Turbine Conference & Products Show, Mrch 9-12, 1981, Houston, Texs. Mnuscript received t ASM Hedqurters December 11, 198. As will be shown, this reltionship cn be used to predict the chnge in N for chnges in humidity, fuel type, power level nd diluent injection. Downloded From: on 11/11/218 Terms of Use: Copies will be vilble until December 1, 1981.

2 1 FIGUR 1 26 FIGUR 2 Dt from Roll 259 Dry Air n Z Cl N, = 7-5 x 1 h e'" 1 11 " C Air With.2 Specific Humidity Temperture - K Fuel/Air Rtio - F/A HUMIDITY FFCT Air contining wter vpor will produce lower temperture products of combustion thn completely dry ir. This is due to two fundmentl effects: (1) the specific het of wter vpor is lmost twice tht of ir, nd (2) the wter vpor rects with crbon t equivlence rtios greter thn bout.8 to form crbon monoxide nd hydrogen, nd bsorbs het in the process. Fig. 2 presents the theoreticl flme tempertures for typicl hydrocrbon fuel burned with dry ir nd with ir contining 2", by weight (.2 specific humidity) of wter vpor. As the figure shows, t the pek flme temperture, where the combustion occurs in diffusion flmes or droplet burning systems, the ddition of 2"o wter vpor hs reduced the temperture 46 K. Thus, in the norml rnge of humidities, the effect of humidity on pek flme temperture cn be expressed by eqution : AT = (46/.2) (S.H.) = (2) (S.H.) () where S.H. is the weight rtio of wter/ir commonly known s specific humidity. This reltionship is vlid for wter concentrtions in ir up to t lest 6 o which covers the rnge of interest in gs turbines. If the temperture difference due to humidity from eqution is inserted in eqution 2, the result is eqution 4: N. Rtio = e(8.28x1-)(2)(s.h.) = e(19)(s.h.) (4) This is the sme effect of humidity on N. s hs been empiriclly determined for gs turbine engines nd promulgted in the PA regultions on N. emissions from sttionry sources. It should be noted tht the ddition of wter vpor to the ir not only lowered the pek flme temperture ner the stoichiometric fuel-ir rtio but lso lowered the turbine inlet temperture nd reduced power output of the engine. In most engines, the fuel control will increse fuel flow to restore power. In combustion systems which burn fuel drops (or in gseous diffusion flmes) this results in more fuel being burned t the lower temperture with negligible effect on N, from the slightly incresed fuel flow. In len, homogenous combustion systems, however, the combustion occurs fr below the pek flme temperture nd the incresed fuel flow restores the flme temperture to the sme vlue it hd with dry ir. For this reson, inlet ir humidity hs no effect on the NX emissions from combustors tht burn len, homogenous fuel-vpor ir mixture. FUL TYP For droplet nd diffusion flme combustion systems, different fuels produce different pek (ner stoichiometric) flme tempertures. The following tble presents the lower heting vlue nd pek flme tempertures for vrious fuels t 15 tmospheres pressure nd n inlet ir temperture of 7 K. Fuel LHV Kg Cl/g Pek Temperture K Methne thne Propne I-Butne I-Pentne Hexne Methnol* thnol No. 2 il *Commercil grde with o wter. 2 Downloded From: on 11/11/218 Terms of Use:

3 The difference in pek flme temperture cn be used with eqution 2 s the first step to predict the rtio of the N. produced by two different fuels. Another fctor lso ffects the mount of N5 formed nd tht is the quntity of fuel burned in given mount of ir. If, for exmple, the N x is formed in stoichiometric diffusion flme surrounding fuel droplet, it is resonble to expect tht incresing the number of drops would increse the mount of N. generted. It ws found empiriclly tht the N. increses s the two-thirds power of the rtio of fuel mss burned. Therefore, eqution 5 cn be used to predict the chnge in N, t generted by gs turbine engine when only the fuel is chnged. N x2 = [N.,] [e(8.28x1 - )(T2-T1)][W f2/w 112/ (5) where T is the pek flme temperture, nd W f is the fuel flow rte which is inversely proportionl to the fuel heting vlue. No fundmentl explntion for the use of the 2/ exponent on fuel flow rtio hs been discovered. It is purely empiricl. The vlidity of eqution 5 is indicted by Figs. nd 4. Fig. compres the mesured N, emissions with the vlues predicted from eqution 5. To mke the predictions, mesured N., fuel heting vlue nd pek flme tempertures for the sme engines operting t the sme power conditions on number 2 oil were used in eqution 5 with the theoreticl flme tempertures nd heting vlues for methnol. Fig. 4 is similr plot for nturl gs. Although the vilble dt where two fuels were run in the sme engine re scrce, the greement between the predicted nd mesured vlues is very good. FIGUR emissions mesured on four FT4-C gs turbine engines compred with vlues predicted by eqution 6. In mking the clcultions, the mesured vlue of 166 PPMV ws used s N., in the eqution. In using these equtions it is importnt not to confuse pek flme temperture with turbine inlet temperture. As power level is chnged, turbine inlet temperture chnges substntilly. The pek flme temperture in the ner-stoichiometric combustion zone surrounding ech fuel drop, however, chnges by much smller mount. For moderte chnges in pressure, the chnge in pek flme temperture cn be closely pproximted s one-hlf the chnge in burner inlet ir temperture z ) N. 4 FIGUR 4 Nturl Gs Dt from Ref 5 / _ 6 Methnol Dt from Ref 4 Line of qulity 5 Q Dt from Ref 6 C^ / Mesured N, - ppmv Z N N. 2C Q ine of qulity 2 18 Sym ngine S/N P6867 P P Q P6867 FIGUR PWR LVL Mesured N, - ppmv When the power level t which n engine is operting is chnged, three prmeters ffect the mount of N. generted. These re (1) the pressure level, (2) the mount of fuel burned, nd () the burner inlet ir temperture which ffects the pek flme temperture. If the N, t one operting level is known, the N x t ny other level cn be predicted from eqution 6: N 52 = [Nxt] [e(8.28x1 )(T2 T1)][ W/W f1 12/ [P 2/P 1] 1 / 2. (6) It is obvious tht eqution 6 is identicl to eqution 5 except for term to djust the N 5 for the chnge in combustion pressure (P) resulting from the chnge in power level. Fig. 5 is comprison of Z U Power Level Dt from Ref 4 Line of qulity Mesured N, - ppmv Downloded From: on 11/11/218 Terms of Use:

4 DILUNT INJCTIN Wter nd stem re common diluents used to reduce N emissions from gs turbine engines. The sme principles discussed bove cn be used to predict the results of diluent injection. The kinetics nd specific het effects of the wter vpor on the pek flme temperture hve lredy been discussed. The other fctor which must be considered is the effect of the diluent on the burner inlet ir temperture. Wter hs lrge het of vporiztion nd is injected t temperture much lower thn burner inlet ir temperture. The totl enthlpy chnge from liquid wter to stem in equilibrium with the burner ir must be considered. The mount of het required to vporize the wter nd het it to the equilibrium temperture of the wter-vpor ir mixture depends on the mount of wter involved. The mount of wter is usully expressed in terms of wter/fuel rtio (W/F) nd the wter is usully mixed with the fuel or injected djcent to the fuel. If the fuel drops burn ner stoichiometric, the mount of wter involved in the combustion process round ech fuel drop cn be expressed s the product of the wter/fuel rtio nd the stoichiometric fuel/ir rtio. It hs been found empiriclly tht fctor of.9 must be included to mke the clculted dt fit the experimentl dt. Two possible explntions for the.9 fctor re (1) diffusion of wter wy from the flme front or (2) combustion round ech drop tkes plce t fuel-ir rtio little lener thn stoichiometric. Whtever the reson, the empiriclly determined wter/ir rtio is expressed by eqution 7: W = (W/F) (F/A,) (.9) (7) A where F/A s is the stoichiometric fuel-ir rtio for the fuel used. The wter/ir rtio is used to clculte the reduction in the temperture of the ir entering the combustion process so tht temperture, in turn, cn be used to clculte the resulting reduction in pek flme temperture. The ctul het bsorbed by the wter in vporizing it nd rising its temperture comes from cooling the ir. Stem tbles nd ir tbles cn be used to rrive t the equilibrium temperture of the wter-vpor ir mixture but it is tril nd error process. The equilibrium temperture cn be very closely pproximted for gs turbine systems by eqution 8: the figure. Although the sctter in the mesurements is lrge, this is not uncommon for N x mesurements t utility instlltions. The predicted dt represent the verge of the experimentl dt very well. The sme pproch cn be used for stem except tht the equilibrium ir temperture TM must be clculted from eqution 12 which hs the ltent het of vporiztion removed nd includes new term, Ts, the temperture of the stem t injection z. Z.4, -.2 T (.24) (TG) + (.46) (Ts) (W/A) M (.46) (W/A) + (.24) Figure 7 shows how the vilble dt compre with the predicted dt bsed on n ssumed stem temperture of 48 K. FIGUR 6 Wter Injection Ref Q UTC FT4 t Fl, No. 2 il 4 UTC FT4 t Conn, No. 2 il 4 Q G 71 B. No. 2 CI 7 Q Undefined ngine. Nt Gs 8 X Undefined ngine, No. 2 il 8 X (12) ` Line of qulity (.24) (TG) (442) (W/A) (8) TM (.46) (W/A) + (.24) where T M is the mixture equilibrium temperture nd T G is the burner inlet temperture before wter injection. Tempertures re ll in K. The reduction in the temperture of the ir tht enters into the combustion process, TC, is the difference between T G nd TM ltc = TM TG. (9) For most hydrocrbon fuels the chnge in pek flme temperture is lmost exctly one-hlf the chnge in inlet ir temperture. ^TF = (.5) (AT). (1) qution 1 defines the chnge in pek flme temperture cused by the wter's cooling effect on the burner inlet ir, qution hs lredy defined the dditionl effect of wter vpor on pek fme temperture. qutions 1, 7 nd cn he combined in eqution 11 to define the totl effect of wter injection on flme temperture. AT = ATF + (2) (W/A). (11) The temperture difference (AT) defined in eqution 11 cn then be used in eqution 2 to predict the effect of wter injection on N X emissions. Fig. 6 shows dt from five different sources. T G ws not vilble for ll the dt so vlue of 6 K ws ssumed to clculte the dt predicted from equtions 11 nd 2 nd shown on 'Pt 1. C z. z.6 z 12 4 d Mesured N, Rtio - N, (Wet)/N, (Dry) FIGUR 7 Stem Injection Dt from Ref 4, Site No. 1 Dt from Ref 4, Site No 2 Dt from Ref 4. Site No. Q Line of qulity Mesured N, Rtio - N,(Wet)/N(Dry) 4 Downloded From: on 11/11/218 Terms of Use:

5 CNCLUSIN Within the sctter of vilble engine dt, the correltions developed in this report relibly predict the N. emissions from industril engines s function of fuel type, power level nd diluent injection. RFRNCS 1. Anderson, Dvid, "ffects of quivlence Rtio nd Dwell Time on xhust missions From n xperimentl Premixing, Prevporizing Burner," NASA TM X-71592, Hrris, Mrgret., et l., "Reduction of Air Pollutnts from Gs Burner Flmes," U.S. Dept. of Interior, Bureu of Mines, Bulletin 65, Lewis, George D., et l., "Design of Successful Low missions Burner," ASM 8-GT-12, Privte Communictions United Technologies Corportion FT4 C Gs Turbine ngines, missions Mesurements t Vrious Utilities. 5. Crl, D.., et l., "xhust missions from 25 MW Gs Turbine Firing Hevy nd Light Distillte Fuel ils nd Nturl Gs," Westinghouse lectric Corp., ASM 75-GT-68, Klptch, R. D., "Gs Turbine missions nd Performnce on Methnol Fuel," ASM 75-PWR 22, Lrkin, R. nd Higgenbothm,. B. "Field Testing of Simple Cycle Sttionry Utility Gs Turbine ffects of Wter Injection for N, Control on missions nd Unit pertions," Acurex Corportion Report 79-58, Durkee, K. R., et l., "Stndrd Support nd nvironmentl Impct Sttement An Investigtion of the Best Systems of mission Reduction for Sttionry Gs Turbines," U.S. nvironmentl Protection Agency, Downloded From: on 11/11/218 Terms of Use: 5

631 ALLOWANES AND TOLERANCES

631 ALLOWANES AND TOLERANCES ANSI Fits. Tbles 3 through 9 inclusive show series of stndrd types nd clsses of fits on unilterl hole bsis, such tht the fit produced by mting prts in ny one clss will produce