| 摘要: | 此論文透過單和雙通道式放電，實驗研究層流和紊流預混球狀火焰的引燃問題。使用傳統火花放電(CSSD，以字母C代表)和奈秒重複脈衝放電(NRPD，以字母N代表)兩種引燃系統，進行引燃能力的比較。所有的電極探針以端面為平面之不鏽鋼製成，其直徑為1 mm。有關雙通道式放電配置，由兩支間距為
2.6 mm的平行陽極電極和一支置於兩陽極電極中間對面且在同一垂直平面上的陰極所組成，此配置能夠形成兩個空間上分離且時間上同步的引燃通道。實驗使用當量比 = 0.7，且有效路易斯數Le 2.2 ≫ 1的貧油正丁烷/空氣混合燃氣，在一雙腔體等溫等壓風扇擾動爆炸設備中進行，此設備可產生一近似等向性紊流場，本研究採用一範圍之均方根紊流擾動速度(u = 0-4.9 m/s)來進行實驗，其平均流速可忽略不計。對於dgap = 0.8 mm的單通道和雙通道的傳統火花電極配置(CS-0.8和CD-0.8-0對應CS-dgap和CD-dgap-L，其中L為陽極和陰極電極端面間之橫向距離於同一垂直平面)。我們發現在靜止流場時，CS-0.8和CD-0.8-0具有相同的引燃特性，這可以從50%引燃機率之相同層流最小引燃能量(MIEL
18 mJ)、臨界火焰半徑(Rc 3.7 mm)以及初始火核發展延遲時間(delay 11.3 ms)看出。在不同的dgap 值，NS-dgap使用9個序列脈衝，ND-dgap-L則使用6個序列脈衝，來達到與CS-0.8和CD-0.8-0具有相同的引燃能量Etot 18 mJ MIEL，其中除NS和ND第一個脈衝有相同的引燃能量Eig,1st 0.8 mJ，NS每個脈衝引燃能量Eig 2.2 mJ，而ND每個脈衝引燃能量Eig 3.3 mJ，且NS與ND配置之引燃能量不受dgap和重複脈衝頻率(PRF)所影響。層流與紊流之引燃機率(Pig,L 和 Pig,T)在PRF = 1-100 kHz和u = 0-4.9 m/s之條件進行量測。我們發現ND-0.8-0之配置，在所有PRF值，其Pig,L = 0%，這與普遍認知的引燃火核越大，Pig,L值越高的觀點不同。但在NS-0.8之條件，在PRF = 40 kHz時，會出現加成效應，歸因於因PRF與向內流動之反應物迴流頻率(fRC)耦合，其Pig,L = 58% > 50%，使用相同Etot MIEL(CS-0.8) 18 mJ作比較。但當dgap = 2.0 mm則有相反結果，在
PRFs = 1-5 kHz時，Pig,L(ND) > Pig,L(NS);在PRFs = 10-100 kHz，則Pig,L(ND) = Pig,L(NS) = 100%。對於PRF = 40 kHz和dgap = 2.0 mm之紊流條件，顯著地觀察到有三個引燃區域具有截然不同的Pig,T：(I) 當u < 1 m/s時，Pig,T(ND) = Pig,T(NS) = 100%;(II) 當1 m/s < u < 4 m/s時，Pig,T(ND) > Pig,T(NS)，顯示ND配置具有引燃增強效果;(III) 當u > 4.2 m/s時，則Pig,T(ND) = Pig,T(NS) = 0%。與NS-2.0和/或ND-2.0-1.8之配置相比，ND 配置的dgap進一步增加到2.8 mm會使Pig,L(ND)和Pig,T(ND)惡化。最後，透過分析CS/CD和/或NS/ND之火核發展高速紋影影像，來解釋前述實驗結果，本研究有助於擬定貧油火花引燃引擎之引燃策略。
;In this thesis, single- and dual-channel sparks are used to investigate experimentally laminar and turbulent premixed spherical flame initiation using both conventional-single-shot-discharge (CSSD represented by the Alphabet C) and nanosecond-repetition-pulse-discharge (NRPD represented by the Alphabet N) ignition systems, and their ignition abilities are compared. All electrodes are made of stainless steel and 1 mm in diameter with flat-end. The dual-channel spark configuration is consisted of two parallel anode electrodes with a fixed apart distance of 2.6 mm and one central cathode electrode on the same vertical plane, capable of forming two spatially separated and temporally synchronized spark channels. Ignition experiments using the lean n-butane/air mixture at the equivalent ratio = 0.7 with an effective Lewis number Le 2.2 ≫ 1 are conducted in a double-chamber, constant-temperature/pressure, fan-stirred explosion facility that can be used to generate a near-isotropic turbulent flow field with negligible mean velocities over a wide range of r.m.s turbulent fluctuating velocity (u = 0-4.9 m/s). For conventional single- and dual-channel electrode configurations at spark gap dgap = 0.8 mm (CS-0.8 and CD-0.8-0, corresponding to
CS-dgap and CD-dgap-L where L is the tip distance between anode and cathode electrodes on the same vertical plane), nearly the same ignition characteristics in quiescence are found, as can be seen by nearly the same values of laminar minimum ignition energy at 50% ignitability (MIEL 18 mJ), critical flame radius (Rc 3.7 mm), and initial kernel development delay time (delay 11.3 ms). The same total ignition energy Etot 18 mJ
MIEL for CS-0.8 and CD-0.8-0 cases is applied to NS-dgap and ND-dgap-L cases at different values of dgap via a train of 9 pulses (NS) and 6 pulses (ND), where each pulse has an ignition energy Eig 2.2 mJ (NS) and 3.3 mJ (ND) except for the first pulse having smaller but the same Eig,1st 0.8 mJ for both NS and ND configurations regardless of dgap and the pulse repetition frequency (PRF). Laminar and turbulent ignition probabilities (Pig,L and Pig,T) are measured over a range of PRF = 1-100 kHz and u = 0-4.9 m/s. Unlike the commonly-held view that the larger the ignition kernel is, the higher the value of Pig,L is, we find no successful ignition for the case of ND-0.8-0, where Pig,L = 0% at all values of PRF. But in the NS-0.8 case, there is a synergistic effect due to the coherence between PRF and the inward reactant flow recirculation frequency (fRC) at PRF = 40 kHz, where Pig,L = 58% > 50% at the same Etot MIEL(CS-0.8) 18 mJ. The result is reversed when dgap = 2.0 mm, where Pig,L(ND) > Pig,L(NS) at PRFs = 1-5 kHz and Pig,L(ND) = Pig,L(NS) = 100% for PRFs = 10-100 kHz. As to the turbulent cases at
PRF = 40 kHz and dgap = 2.0 mm, three ignition regimes with drastically different Pig,T are observed: (I) Pig,T(ND) = Pig,T(NS) = 100% when u < 1 m/s; (II) Pig,T(ND) > Pig,T(NS) when 1 m/s < u < 4 m/s showing the ignition enhancement of ND configuration;
(III) Pig,T(ND) = Pig,T(NS) = 0% when u > 4.2 m/s. A further increase of dgap to 2.8 mm for the ND configuration deteriorates Pig,L(ND) and Pig,T(ND) as compared to those of NS-2.0 and/or ND-2.0-1.8 configurations. Finally, high-speed schlieren images of CS/CD and/or NS/ND kernel evolutions are shown to comprehend these results which may be relevant to a better ignition strategy selection in lean-burn spark ignition engines. |
