單Tsuji燃燒器上於火焰轉換過程中之火焰行為的數值研究

Abstract

本篇論文已經完成單Tsuji燃燒器上二維火焰的數值模擬去研究相關火焰的特性與控制機制，研究的案例包含穩態的案例與暫態的案例。在穩態的案例中，它首先使用甲烷/氮氣為燃料去研究氮的稀釋效應，然後它使用甲烷/空氣燃料去研究空氣的預混效應。在暫態的案例中，它研究火焰行為的浮力效應以及輻射效應，所採用的燃料是純甲烷。從這些數值結果中發現：對於甲烷/氮氣燃料，在甲烷質量分率低於0.6的區域，當進氣速度逐漸增加至極限值時，火焰會由包封火焰直接轉變成為尾焰。而在甲烷質量分率高於0.6的區域，當進氣速度增加時，昇離火焰將出現在包封焰和尾焰之間。除此之外，當甲烷質量分率降低時，火焰轉換的極限速度也隨之降低，這預測的趨勢與在2005年，蔡耀慶同學實驗觀測的結果，有定性的相似。對於甲烷/空氣燃料案例中，本研究發現火焰轉換的順序是和甲烷/氮氣燃料火焰相同。然而，包封火焰不再是單純的擴散焰。它的火焰的結構已被Makino等人在1991年的實驗確認。除此之外，在火焰轉換區中的昇離火焰會產生回火現象。對於甲烷質量分率大於0.7的甲烷/空氣燃料，火焰轉換極限速度幾乎是不會隨著甲烷的質量分率改變。在暫態的案例中，在進氣速度為0.6 m/s的冷流流場結構，會在風洞內看見兩個渦流同時存在的現象。這現象會一再出現，直到4.71秒為止。在此之後，一個渦流將穩定的存在於燃燒器的後方。當燃料從燃燒器的前半部以0.05m/s的噴氣速度朝著進氣流的方向噴出時，產生的流場幾乎是不會隨著時間改變。因此，它可視為一個穩定流場。當氣體點燃時，因為活躍的化學反應，大量的熱被釋放，導致氣體溫度快速增加，並產生一個突然的氣體膨脹。因此，進氣流被氣體膨脹產生的逆向壓力所阻滯，火焰沿著燃燒器向上、下游延燒。在點火後0.03秒的時候，預混氣體燒盡。隨著時間逐漸進行，氣體膨脹所產生的效應減少，火焰將逐漸往後退，然後發展成一個沒有渦流的包封擴散火焰。因為過多的燃料，這火焰呈現動盪的特性，火焰尾部的燃料快速消耗。在點火後0.48秒的時候，原先在點火過程中所聚集的燃料燒完。此時，火焰會產生局部熄火的現象，然後，它將藉由反應區的高溫復燃。它是一個較弱的火焰。雖然，在點火後0.48秒的時候，渦流再次形成在燃燒器後方，但是，因微弱火焰所產生的氣體膨脹無法維持這渦流。因此，它將會再次消失。隨著時間的增加，火焰會因燃燒器前半部噴射的燃料使火焰再次變強。當時間到達點火後0.74秒的時候，渦流再次形成在燃燒器後方，且不再消失。然而，因先前的擾動，這火焰仍然是不穩定。一直到時間為點火後2.2秒的時候，火焰才變成穩定。當火焰變為穩定的包封火焰狀態時，進氣流將被調整到設定值，兩個火焰吹離的機制(冷卻效應與火焰拉伸)被確認。當進氣速度增加至極限值時，火焰會被拉斷，在此之後，反應區顯示三種火焰結構，它們分別是純預混焰、過度區以及純擴散焰。在考慮浮力效應的情況下，預測的火焰形狀與停滯距離是符合2005年蔡耀慶同學實驗觀測的結果。此外，昇離火焰也將短暫出現在包封火焰和尾焰之間，然後穩定變成尾焰。藉由無因次化過程的證明，氣態輻射能被忽略不計。然而，固態輻射卻必須納入考量。

This dissertation has completed the numerical simulation for a two-dimensional flame over a Tsuji burner to investigate the corresponding flame characteristics and the controlling mechanism. The investigation includes the steady and unsteady cases. In the steady cases, they first simulate Tsai (2005) experiment to investigate the dilution effect of nitrogen by using methane/nitrogen mixture as the fuel. The other changes the fuel to methane/air mixture that aims to investigate its premixed effect. In unsteady cases, they investigate the buoyancy and radiation effects on the flame behaviors. The adopted fuel is pure methane. The numerical results find that for the steady cases of methane/nitrogen fuel, in the regime of , the envelope flame will transform into wake one directly when incoming velocity gradually increases to a limiting value. On the other hand, in the regime of , the lift-off flame will appear between the envelope and wake ones as the velocity increases. The flame blow-off limits and critical transition velocity are reduced as the mass fraction of methane in fuel mixture is lowered. The predicted trends are similar to the ones observed experimentally in Tsai (2005). For methane/air mixture, the flame transition order is same as the one for methane/nitrogen mixture. However, the envelope flame is not a pure diffusion flame, whose structure has been verified experimentally by Makino et al. (1991). The lift-off flame during flame transition regime will generate the flashback phenomenon. The blow-off limits and critical transition velocities are almost invariant with when is greater than 0.7. In the unsteady cases, it is found that as inflow velocity is fixed at Uin=0.6 m/s, the cold flow structure can observe the co-existence of two recirculation flows in the wind tunnel. The flow variations repeat to occur until t = 4.71 sec. After that, a vortex is stably existed behind the burner and no shading phenomenon occurs. When the fuel, methane, is ejected from the forward half part of a Tsuji burner with Vw = 0.05m/s into the incoming air flow, the resultant flows are almost invariant with time. Hence, it can be regarded as a stable flowfield to exist. As the flammable mixture is ignited, an active chemical reaction occurs along the cylindrical burner that leads to a rapid rise in gas temperature and a sudden gas expansion. Meanwhile, this incoming flow ahead of the flame front is retarded due to the adverse pressure. The flame spreads toward the upstream and downstream along the burner. At the time, t=0.03 sec. after ignition, the premixed mixture in the flame front is almost burned out. As the time proceeds, the effect of gas thermal expansion is gradually decreased. The flame will gradually retreat and then becomes an envelope flame without recirculation flow. The flame shows an oscillating feature and the fuel in the wake consumes rapidly. As the time reaches 0.48 second after ignition, the fuel accumulated during ignition process is almost burned out and the flame will generate the local extinctions, then, it will be reignited. At this moment, the recirculation flow reappears behind the burner, but later it will disappear rapidly. After that, the flame will gradually become stronger again. As the time reaches 0.74 second after ignition, the recirculation flow behind the cylindrical burner reappears and it will not disappear. However, the flame is still unstable because of the previous perturbation and the complicated interaction between the chemical reaction and gas expansion until t=2.2 sec. after ignition. As it reaches a stable enveloped flame condition, the inflow velocity, Uin, will be adjusted to an assigned value. The two blow-off mechanisms, quenching effect and flame stretch, are identified. As inflow velocity increases to a limit, the reaction zone will be cut off. After that, the reaction zone behind the surface burner shows three flame structures, which consist of the pure premixed flame, the flame which shows diffusion and premixed flame characteristics together, and the pure diffusion flame. The predicted unsteady results with buoyancy effect agree well with the ones observed experimentally in Tsai (2005) for the flame shapes and stand-off distance. Additionally, the lift-off flame will appear briefly between the envelope and wake ones, then, is stabilized as wake flame. The gas phase radiation effect can be neglected via the argument of normalization procedure, but the consideration of solid phase radiation is necessary.