摘要: | 本論文定量量測汽油主要參考燃料(Primary Reference Fuel, PRF) 於高溫、高壓條件下之層、紊流燃燒速度(SL、ST),使用體積比為95%異辛烷與5%正庚烷之PRF95預混燃氣。主要研究重點有二:(1)量測液態燃料PRF95球狀火焰之SL和ST,探討在固定方均根紊流擾動速度(u′)條件下,壓力(p)效應對於SL和ST之影響;(2)探討在固定流場紊流雷諾數ReT,flow = u′LI/條件下,壓力效應與ReT,flow對ST之影響,其中LI為紊流積分長度尺度和為運動黏滯係數。實驗在已建立之高溫高壓高紊流強度雙腔體預混燃燒設備中進行,透過內爐十字型燃燒器大水平圓管內兩端之一對10匹馬力馬達所驅動之反向旋轉風扇和空孔板,於十字型燃燒器中心引燃區產生一近似等向性紊流場。因會隨著p增加而下降,故我們在高壓條件下,等比例地調控u′LI之乘積使其與下降幅度相同,以獲得固定ReT,flow值。實驗條件如下:化學計量PRF95燃氣(當量比 = 1.0;有效Lewis數Le ≈ 1.44)、溫度T = 358K和373K、p =1-5 atm、u′ = 1.0-5.4 m/s,以及固定三個不同ReT,flow值(6,700;9,100;11,600),以探討不同p與ReT,flow效應對ST之影響。對於層焰而言,p增加會使SL下降,這是由於壓力效應對於層焰傳遞過程中為一個壓力連鎖反應機制,以H-O2反應為例,關鍵反應式(H1):H + O2 → OH + O為二體分支反應;而反應式(H9):H + O2 + M → HO2 + M為三體抑制反應,當壓力增加時,三體反應較二體反應強烈,因此在高壓條件的整體反應過程中,抑制反應被加強導致SL下降。對於紊焰而言,當u′值固定時,ST值會隨壓力增加而增加,過往認知這是因為火焰厚度會隨壓力增加而減少,使高壓火焰表面易受流力不穩定性(Hydrodynamic Instability,也就是Darrieus-Landau Instability)之影響,火焰表面會產生許多細胞狀結構(Cellular Structure),造成紊焰面積增加而使得ST增加,但過往研究未考慮當p增加會使ReT,flow增加之事實。實驗結果顯示,在固定u′條件下,ST值會隨壓力增加而上升,這很可能是由於因p上升會使得下降進而使ReT,flow增加,致使ST值增加。然而,當在固定ReT,flow條件下,有別於以往認知,ST值則是會隨著p增加而下降,且其趨勢與SL相似,即ST和SL對壓力效應均呈指數下降之關係,這顯示ReT,flow可能是影響ST增加的重要因素。 本論文亦整合先前實驗室之燃燒速度數據,以常用之五組一般通式進行正規化分析:(1) ST,c=0.5/u′ = A(Da)B,其中Damköhler number Da = (LI/u′)(SL/L)、c ̅為平均傳遞變數、L為層流火焰厚度;(2) ST,c=0.5/SL = A(ReT,flame)0.5,其中火焰紊流雷諾數ReT,flame= (u′/SL)(<R>/L)、<R>為平均火焰半徑;(3) ST,c=0.5/SL = A[(u′/SL)(p/p0)]B,其中p0為一大氣壓;(4) ST/u′ = A K B,其中Karlovitz number K = 0.157(u′/SL)2(ReT,flow)-0.5;(5)ST,c=0.5/SL = A(ReT,flame/Le2)B,其中A、B分別為不同一般通式之實驗常數。結果發現,原先分散之正規化ST/SL數據,代入經考量Le數為修正參數之函數後,可收斂合併成單一曲線,五組修正後之一般通式皆展現很好的藕合程度,其中以一般通式(3)有最佳的藕合程度可達(goodness, R2 = 0.90)。結果顯示火焰傳播即使在不同燃料與當量比條件下,依然具有自我傳播相似性。本論文高溫高壓預混紊流燃燒之研究結果,對汽車引擎燃燒研究應有所助益。 ;This thesis quantitatively measures laminar and turbulent burning velocities (SL and ST) of the premixed PRF95 (primary reference fuel with 95% volume of iso-octane and 5% volume of n-heptane) at high temperature and high pressure conditions. There are two objectives in this thesis: (1) To measure values of SL and ST of spherical premixed flames of PRF95 and investigate the pressure effect on SL and ST by keeping the r.m.s. turbulent fluctuation velocity (u′) constant. (2) To investigate the effect of pressure and turbulent flow Reynolds number (ReT,flow = u′LI/) on ST under constant ReT,flow condition, where LI and are the integral length scale of turbulence and the kinematic viscosity of reactants respectively. Experiments are conducted in a large dual-chamber, high-pressure, high-temperature, fan-stirred cruciform premixed turbulent explosion facility capable of generating isotropic turbulence for ST measurements of expanding spherical PRF95/air flames. While controlling the product of u′LI in proportion to the decreasing at elevated pressure to obtain the constant ReT,flow value, and we measure SL and ST of PRF95 at the equivalence ratio ( = 1.0, Le ≈ 1.44), T = 358K, and 373K, p = 1-5 atm, u′= 1.0-5.4 m/s, and at three constant ReT,flow = 6,700; 9,100; 11,600 in order to further investigate the ReT,flow effect on ST. Under high pressure condition, SL decreases with increasing pressure, where cellular structures all over the flame surface can be observed. For laminar flames, SL decreases with increasing pressure because SL is effected by a pressure chain reaction mechanism during flame propagating. Take H-O2 reaction for example, key reaction (H1): H + O2 → OH + O is a two-body branching reaction, while reaction (H9): H + O2 + M → HO2 + M is a three-body inhibiting reaction. While by increasing pressure, (H9) is enhanced relative to (H1) because three-body reactions are favored over two-body reactions as pressure increases. Therefore, in the overall reaction at high pressure conditions, the inhibition reaction is enhanced to cause a decrease in SL. For turbulence flames, when under constant u′ condition, ST increases with increasing pressure. In the past, this is because the flame thickness decreases with increasing pressure, making the flame surface interfered with Hydrodynamic Instability (i.e., Darrieus-Landau Instability). Flame surface results in the cellular structures which causes an increase in the area of the turbulent flame and increases the ST. However, previous studies have not considered the fact that pressure increases with the ReT,flow increasing. Results show that when under the condition of constant u′, the values of ST increase with increasing p, which is mainly due to increase of ReT,flow with increasing p. On the contrary, when under the condition of constant ReT,flow, we find that the values of ST actually decrease with increasing p. Both ST and SL have an exponential decrease in pressure effect which indicates that the ReT,flow may be a key factor affecting the increase of ST. This thesis also includes previous iso-octane ST at T=358K, 373K, and 423K obtained from data our laboratory using the same methodology, so that the normalization of these data by using five general correlations can be made for comparison as describe below (1) ST,c=0.5/u′ = A(Da)B, where the subscript c is the mean progress variable, Damköhler number Da = (LI/u′)(SL/L), L is laminar flame thickness. (2) ST,c=0.5/SL = A(ReT,flame)0.5, where the turbulent flame Reynolds number ReT,flame= (u′/SL)(<R>/L). (3)ST,c=0.5/SL = A[(u′/SL)(p/p0)]B, where p0 is the atmospheric pressure. (4)ST/u′=AK-0.3, where Karlovitz number K = 0.157(u′/SL)2(ReT,flow)-0.5. (5)ST,c=0.5/SL= A(ReT,flame/Le2)B and A and B are experimental constants of different general correlations. As a result, it is found that the previously dispersed normalized ST/SL data can be converged into a single curve after being substituted into a function that considers the Le number as a correction. The five modified general correlations all show good fitting goodness. Among them, the general correlation (3) has better fitting goodness (R2 = 0.90). The results show that the flame propagation still has self-propagation similarity even though for different fuel and equivalence ratios. The results of this thesis are useful for high temperature and high pressure premixed turbulent combustion applications, such as combustion studies for vehicles engines. |