摘要: | 本論文探討壓力效應對於中央引燃預混火焰傳播速度(flame propagation speed, SF)之影響。本研究使用低碳燃料甲烷和空氣之預混燃氣,針對當量比(equivalence ratio)?? = 0.8、1.0與1.2,在0.1 ~ 1 MPa等不同壓力條件下,於高壓十字型預混燃燒器中,執行一系列中央引燃之層流與紊流燃燒實驗。高壓十字型預混燃燒器含兩大部分,一為大型高壓保護外爐,另一為置於其內之十字型燃燒器。內外爐皆具有四個耐高壓之石英玻璃視窗,配合高速、高解析度CMOS攝影機搭配上濾光片透過視窗,可用來觀測火核與火焰的成長以及與流場的交相作用。十字型燃燒器內的紊流場由置於其水平圓管兩端之一對反向旋轉特製風扇及空孔板所產生,可在燃燒器中央觀測區產生一約15公分立方具零平均速度之等向性紊流場,其紊流擾動速度(u')最高約可達8 m/s,相對應之ReT = u'LI/? = 24,850,其中LI與?分別為流場積分長度尺度和流體運動黏滯係數。預混火焰熱擴散不穩定效應,可用無因次參數Lewis(Le)或Markstein (Ma)數來說明,其中Le = αT/D,αT和D分別為熱與質量擴散係數;而層流Ma = LM /?F,LM為Markstein長度,?F為火焰厚度(flame thickness)。由貧油實驗結果顯示,在常壓(0.1 MPa)條件下,當Le < 1或Ma數較小、接近於0和為負值時,貧油預混甲烷火焰(? = 0.8)其未受拉伸層流燃燒速度(SL,0)為0.31 m/s,且火焰面會因熱擴散不穩定,使火焰面產生細胞狀結構(cell structure)。SL,0值會隨著壓力增加而下降,隨著壓力提昇至1 MPa時,SL,0值僅為0.09 m/s,下降約71%,且?F值亦下降,使熱擴散不穩定效應更容易發生,令細胞狀結構更加明顯。在常壓層流富油預混甲烷(? = 1.2)條件下,Le > 1相對應之Ma 1.81,SL,0約為0.34 m/s,火焰面上沒有細胞狀結構。同樣地,SL,0值隨壓力上昇而下降,在1 MPa時,SL,0 ≈ 0.13 m/s,且因?F值變小,使火焰面易產生些為皺褶。我們量測所得之SL,0值與壓力之關係,可以用Metghalchi(1980)所提之關係式來表述,即SL,p = SL,0 (Pb/P0)β,其中SL,p為不同壓力下之層流燃燒速度,Pb為初始壓力而P0為參考壓力(0.1 MPa),β為燃氣當量比的函數。在紊流條件下,針對貧油預混甲烷(? = 0.8)火焰,探討壓力效應對紊流燃燒速度(ST)之影響。實驗結果顯示,ST值似乎會隨壓力增加而增加,此結果與層流結果正好相反。且在高壓紊流條件下,熱擴散不穩定效應相較於層流條件時,會更加明顯,如初始壓力為0.5 MPa與風扇轉頻f = 30Hz時,正規化之ST值約為0.1 MPa之3.5倍,在1 MPa時,甚至達7倍之多。所大幅增加之ST值,似乎無法僅由熱擴散不穩定效應受高壓和紊流交互作用來解釋,我們猜測流體動力不穩定也扮演一角色,使得高壓紊焰皺褶增加,進而使其ST值大幅提升。這顯示受到壓力的影響,熱擴散不穩定效應與紊流交互作用,可再增加火焰表面積進而使ST值大幅提升。 This thesis measures the effect of elevated pressure on centrally-ignited premixed flame propagation speeds (SF). We study methane-air mixture at three different equivalence ratios (? = 0.8、1.0、1.2), because methane has the lowest C/H ratio among all hydrocarbon fuels. At each value of ???both laminar and turbulent combustion experiments with initial pressure varying from 0.1 ~ 1 MPa are conducted. All combustion experiments using spark ignition are carried out in the centre of a high-pressure cruciform burner. The high-pressure cruciform burner has two major parts: a huge outer high-pressure absorbing safety chamber and a large inner cruciform burner. Both inner and outer chambers have four optically-accessed quartz windows on their top, bottom, front, and back sides. Thus, visualizations of flame kernel formation and its subsequent flame propagation interacting with turbulence can be recorded by a high-speed, high-resolution camcorder. Using a pair of counter-rotating fans and perforated plates equipped to the two ends of the large horizontal vessel, a near-isotropic turbulent flow field having 150 × 150 × 150 mm3 can be generated in the central region of the inner cruciform burner. In it the maximum value of turbulent fluctuating velocities u' can be up to 8 m/s with the corresponding turbulent Reynolds number ReT = LIu'/ν = 24,850, where LI is the integral length scale and ν is kinematic viscosity of reactants. The dimensionless Lewis number (Le = αT/D) and/or Markstein number (Ma = LM /?F) are used to discuss the effect of diffusion-thermal instability, where αT is the thermal diffusivity, D is the molecular diffusivity, LM is Markstein length, and ?F is flame thickness, respectively. Results show that, at?? = 0.8, where Le > 1 and Ma is either very close to zero or becomes negative under laminar condition, the unstretched laminar burning velocity (SL,0) is about 0.31 m/s when P = 0.1 MPa, of which the cellular structure is observed on the outwardly-propagating flame surface. Such cellularity becomes even more obviously when the pressure increases to 1MPa, where SL,0 ≈ 0.09 m/s with a decrease of 71% compared to that at 0.1 MPa. As P increases, ?F decreases making the flame more vulnerable to the diffusion-thermal instability. At ? = 1.2, where Le > 1 and Ma 1.81, the diffusion-thermal instability is not observed, but the outwardly-propagating flame surface can be slightly wrinkle by the hydrodynamic instability where SL,0 is about 0.34 m/s at P = 0.1 MPa. When increasing the initial pressure to 1 MPa, SL,0 drops to 0.13 m/s. And the flame surface appears more wrinkling due to the decrease of ?F. For laminar case, our SL,0 data can be fitted by an empirical relation, SL,p = SL,0(Pb/P0)β, proposed by Metghalchi et al, where SL,p is the laminar burning velocity under different initial pressures, Pb is the initial pressure, P0 is the reference pressure at 0.1 MPa, and β is a function of ?. The response of turbulent burning velocity (ST) with increasing pressure shows an opposite trend. For example, when ? = 0.8 with f = 30Hz, the normalized value of ST at P = 1 MPa is 7 times larger than that at P = 0.1 MPa case, suggesting that the effect of elevated pressure promotes the interactions between turbulence and flame instabilities resulting in an increase of ST. |