摘要: | 本論文使用貧油預混甲烷,於固定當量比(equivalence ratio)條件下,加入體積比10、20、30%之氫氣或氨氣作為燃料,在初始壓力p = 1~5 atm、方均根紊流擾動速度u′ = 0 ~ 4.2 m/s下,量測其球狀火焰層流和紊流燃燒速度(SL和ST)。實驗在已建立之高溫高壓雙腔體三維十字型燃燒設備進行,透過設置於水平圓柱兩端之反向旋轉風扇及空孔板,在三維十字型燃燒室中心區域產生一近似等向性紊流場,並且搭配高速攝影機紀錄中央引燃往外傳播之球狀火焰,以獲得其火焰半徑之時序資料R(t),來估算其燃燒速度。實驗結果主要含三個部分: (1)在常壓下,探討紊流效應對加氫或加氨效應燃燒速度的影響。結果顯示,ST值會隨著氫氣混燒體積比的提升而增加,而氨氣混燒體積比的提升則會造成ST值的下降,且混氫或混氨ST值皆會隨著u′的提升而增加。(2)探討壓力效應(p = 1~5 atm)對於SL和ST之影響。結果顯示,在層流條件下,體積比10~30%氫氣或氨氣混燒之CH4/H2/air或CH4/NH3/air火焰速度變化為SL ~〖 p〗^(〖-n〗_(1 ) )或〖 p〗^(〖-n〗_(2 ) ),其中n1 = 0.52~0.55,而n2 = 0.29~0.36,兩者結果皆顯示SL隨著壓力增加而成一負冪次方下降。在紊流條件下,體積比10~30%氫氣或氨氣混燒之CH4/H2/air或CH4/NH3/air火焰速度變化為ST ~ p^(〖+n〗_3 )或〖 p〗^(〖+n〗_4 ),其中n3 = 0.04~0.16,而n4 = 0.06~0.18,結果顯示混氫或混氨ST皆會隨著壓力上升而成一正冪次方增加。(3)將量測到的SL及ST進行正規化分析,並加入有效Lewis數(Le)之考量,以比較目前文獻上常用五組不同一般通式之適用性。五個不同通式條列如下: (1) Kobayashi et al. (1998)所提出之"S" _"T,c̅=0.5" "/SL"=A[(u′〖/S〗_L)〖(p〖/p〗_0)]〗^B,其中p0為1 atm,A、B為透過實驗係數常數。(2) Bradley et al. (2005)所提出之"S" _"T,c̅=0.5" /u′=A〖[KLe]〗^B,其中下標c ̅為火焰平均傳遞變數、K = 0.25(u′/SL)2(ReT,flow)-0.5為紊流Karlovitz數,而ReT,flow = u′ LI / v ( LI為紊流積分長度和v為反應物運動黏滯係數)。(3) Chaudhuri et al. (2012)所提出之[(1/SLb)(d<R>/dt)] = A(ReT,flame)B,其中SLb為未經密度校正之生成物層流燃燒速度,而ReT,flame= (u′/SL)(<R>/L),其中<R>為平均火焰半徑,L為層流火焰厚度。(4) Shy et al. (2012)所提出之"S" _"T,c̅=0.5" "/" u′= A(Da)B,其中Da = (LI/u′)(SL/L)為紊流"Damk" "o" ̈"hler" 數。(5) Wang et al. (2020)所提出之"S" _"T,c̅=0.5" 〖"/S" 〗_"L" -"1"=A(〖"Re" 〗_"T,flame" 〖"Le" 〗^"-2" )^B。將本實驗所獲得之實驗數據考慮Le數為修正參數之函數後,代入前述五個一般通式後可以得到: (1) "S" _"T,c̅=0.5" "/SL"=2.97〖[(u′/S_L)(p/p_0)〖Le〗^(-1)]〗^0.43;(2) "S" _"T,c̅=0.5" /u′=1.04〖[KLe]〗^(-0.21);(3)〖" S" 〗_"T,c̅=0.5" "/SL"=0.39〖(〖Re〗_(T,flame) 〖Le〗^(-1))〗^0.5;(4)〖" (S" 〗_"T,c̅=0.5" "/u′ ")=0.26(Da〖"Le" 〗^"-1" )^"0.5" ;(5) "S" _"T,c̅=0.5" 〖"/S" 〗_"L" -"1"=0.14(〖"Re" 〗_"T,flame" 〖"Le" 〗^"-2" )^0.6。所有資料經Le數修正後,可將原先相同當量比但不同燃料之實驗數據耦合成單一曲線,使原本之一般通式在修正後可有更廣泛的通用性,結果顯示火焰傳播即使在不同燃料與當量比條件下,是具有自我傳播相似性。前述高壓預混紊流燃燒之研究成果,對燃氣渦輪機和鍋爐等燃燒設計和應用,應有所助益。;This thesis reports measurements of laminar and turbulent burning velocities (SL and ST) of lean premixed methane/air blending with 10-30 vol.% hydrogen and/or ammonia at a fixed equivalence ratio over wide ranges of pressure p = 1~5 atmand r.m.s. turbulent fluctuation velocities u′ = 0~4.2 m/s. Experiments were conducted in an already-established high-pressure dual-chamber 3-D cruciform explosion facility, capable of generating near-isotropic turbulence. The radii of laminar and turbulent spherical flames as a function of time R(t) were recorded by a high-speed camera to determine their associated values of SL and ST. Three main parts of the results are as follows. (1) The effect of u′ on SL and ST of hydrogen and/or ammonia addition at 1 atm is discussed. Values of SL and ST increase with increasing hydrogen blending but decrease with increasing ammonia blending. All ST values increase with increasing u′. (2) The effect of pressure on SL and ST is explored. Results show that SL ~〖 p〗^(〖-n〗_1 ), where n1 = 0.52~0.55 for CH4/H2/air mixtures and n1 = 0.29~0.36 for CH4/NH3/air mixtures. But ST ~〖 p〗^(〖+n〗_2 ), where n2 = 0.04~0.16 for CH4/H2/air mixtures and n2 = 0.06~0.18 for CH4/NH3/air mixtures. (3) Using measured SL and "S" _"T,c̅=0.5" data, five general correlations commonly used in the literature with the consideration of Lewis number (Le) are tested, where the mean progress variablec ̅=0.5are selected. (1) A correlation of "S" _"T,c̅=0.5" "/SL"=A[(u′〖/S〗_L)〖(p〖/p〗_0)]〗^B proposed by Kobayashi et al. (1998), where p0 was 1 atm, and A and B were experimental coefficients. (2) A correlation of "S" _"T,c̅=0.5" /u′=A〖[KLe]〗^B proposed by Bradley et al. (2005), where the turbulent Karlovitz number K = 0.25(u′/SL)2(ReT,flow)-0.5, the flow turbulent Reynolds number ReT,flow = u′ LI / v, and LI and v are the turbulent integral length scale of turbulence and the kinematic viscosity of reactants. (3) A correlation of [(1/SLb)(d<R>/dt)] = A(ReT,flame)B proposed by Chaudhuri et al. (2012). SLb is the laminar burning velocity on the burned side before density correction. ReT,flame= (u′/SL)(<R>/L), where <R> is the average flame radius and L is the laminar flame thickness. (4) A correlation of "S" _"T,c̅=0.5" "/" u′= A(Da)B proposed by Shy et al. (2012), where the turbulent "Damk" "o" ̈"hler" number Da = (LI/u′)(SL/L). (5) A correlation of "S" _"T,c̅=0.5" /SL - 1 = 0.14(ReT,flameLe-2)0.6 proposed by Wang et al. (2020). Applying the present experimental data with the Le modification, these five general correlations are: (1) "S" _"T,c̅=0.5" "/SL"=2.97〖[(u′/S_L)(p/p_0)〖Le〗^(-1)]〗^0.43; (2) "S" _"T,c̅=0.5" /u′=1.04〖[KLe]〗^(-0.21); (3)〖" S" 〗_"T,c̅=0.5" "/SL"=0.39〖(〖Re〗_(T,flame) 〖Le〗^(-1))〗^0.5;(4)〖" (S" 〗_"T,c̅=0.5" "/u′ ")=0.26(Da〖"Le" 〗^"-1" )^"0.5" ; (5) "S" _"T,c̅=0.5" /SL - 1 = 0.14(ReT,flameLe-2)0.6. All five general correlations show that turbulent spherical flames are self-similar regardless of different and fuels applied. These results should be useful to the combustion design and application of gas turbines and boilers. |