淨煤氣化合成氣貧油可燃極限與燃燒速度量測：壓力和紊流效應

Measurements of Lean Flammability Limits and Burning Velocities for Clean Coal Gasification Syngas

董益銍(Yi-chih Dong)

指導教授 ： 施聖洋

國立中央大學/工學院/機械工程學系/碩士(2012年)

摘要

本論文實驗探討淨煤挾帶流氣化合成氣(氫/一氧化碳)之應用燃燒技術，針對不同氫氣/一氧化碳比例合成氣於常壓與高壓條件下，量測其貧油可燃極限，以及層流與紊流燃燒速度(ST和SL)。所有實驗均於本實驗室已建立大型高壓、雙腔體設計之預混燃燒設施平台，該平台由一內爐與一外爐所組成。內爐為本實驗室已發展多年之十字型燃燒器所設計而成，它由水平與垂直鋼管所構成，透過水平腔體底端兩側適合於高壓環境下操作之兩組三相十匹馬力之馬達，反向旋轉左右設之特製風扇並配合空孔板，可於十字型燃燒器中心處區域產生一近似等向性之紊流場。外爐為一大型高壓氣密保護腔體，可於內爐實驗過程中，即中心引燃往外傳播火焰爆炸過程中，吸收由內爐垂直腔體上四個靈敏的釋壓閥所釋放之壓力。利用高速高解析度攝影機，量測中心引燃往外傳播預混火焰之動態傳遞時序影像，經影像分析火焰拉伸率和密度校正計算後，可獲得ST和SL值。
有關層流預混火焰實驗，我們針對貧油條件當量比(??為?0.4 ~ 1.0三種不同比例50%CO/50%H2、65%CO/35%H2和95%CO/5%H2之合成氣燃料，量測其在不同壓力(p)1 ~ 5大氣壓條件下，SL與可燃極限。結果顯示，隨著p增加，不僅SL值降低，並且貧油可燃極限也會擴展。此外，當?值越接近其貧油可燃極限時，SL值對p增加之效應會更加敏感，但當?值接近化學計量比時(??= 1)，則較不敏感。接著依不同合成氣比例的結果來看，隨著氫濃度比例增加，會有更高的SL值，且SL值對p之敏感度會降低，而其貧油可燃極限會有向下擴展之趨勢。
有關紊流預混火焰我們針對(65%CO/35%H2)在??= 0.5時，量測其高壓紊流燃燒速度，壓力範圍從1大氣壓到10大氣壓，特別指出不同於先前研究的地方，即本實驗中心引燃往外傳播之球狀紊焰是在固定紊流雷諾數(ReT ? u?LI/?)條件下所獲得。我們調控方均根紊流擾動速度(u?)與紊流積分長度尺度(LI)，使其乘積與因p值提升而下降的運動黏性係數(?)做等比例調整，以進行固定ReT之實驗量測，其實驗條件有四個不同ReT值:6,700, 9,100, 11,600, 14,200。結果顯示，相反於一般ST隨p增加而上升(p上升，?下降，ReT增加)，在ReT固定條件。我們發現 ST下降趨勢相似於SL，兩者均會隨p增加呈現指數下降之關係，例如ST ~p-0.49在ReT = 14,200。此外，在任一固定p值，正規化紊流燃燒速度(ST/SL)會隨ReT值增加而有顯著增加，顯示ReT之重要性。前述在不同p和ReT之大範圍散佈ST/SL數據，可被聚合成一曲線關係 ST /u? = 0.49Da0.25，其中Da為紊流Damkohler數，此關係式也於Zimont學者提出之理論解ST /u? ~ Da0.25相同，但於過去本實驗團隊針對於貧油甲烷然氣所進行等ReT實驗結果不同，後者ST /u? = 0.14Da0.5，兩者有不同之Da冪次關係。我們將探討壓力、火焰不穩定性與紊流尺度等效應，嘗試解釋前述不同Da冪次之結果。本研究結果對未來應用於高氫比例之合成氣諸如燃氣輪機、工業爐和其他燃燒器將有重要之助益。

關鍵字

層流和紊流燃燒速度 ； 紊流Damkohler數 ； 紊流雷諾數 ； 高壓預混貧油合成氣燃燒 ； 氫氣/一氧化碳合成氣 ； 貧油可燃極限

並列摘要

This thesis studies experimentally high-pressure lean syngas (hydrogen/carbon monoxide) combustion characteristics derived from the clean coal entrained-flow gasifier. Focuses are placed on measurements of lean flammability limits and laminar and turbulent burning velocities (ST and SL) for various compositions of H2/CO mixtures. Experiments are carried out in a large high-pressure, double-chamber (an inner chamber and an outer chamber) premixed turbulent combustion facility. The inner chamber applies the same design of our previous cruciform burner that was constructed by two mutually crossing cylindrical vessels, a horizontal vessel and vertical vessel, forming a cruciform shape. Using two identical frequency-controlled counter-rotating fans and perforated plates at the two ends of the horizontal vessel, an intense near-isotropic turbulent flow field can be generated in the central region of the inner chamber. As to the outer chamber, it is a large high-pressure airtight chamber used to absorb the pressure rise releasing from the inner chamber via four sensitive pressure-releasing valves. These four valves are placed symmetrically in the front and back sides of the vertical vessel. A high-speed, high-resolution CMOS camera system used to directly record flame propagation images and thus flame burning velocities can be determined from image analysis with flame strain rate calculation and the density correction.
Concerning laminar premixed syngas experiments, we measure the lean flammability limits and laminar burning velocities at various equivalent ratios from ??= 0.4 ~ 1.0 with three different syngas fuel compositions, namely 50%CO/50%H2, 65%CO/35%H2 and 95%CO/5%H2, each covering an initial pressure (p) range from 1 atm to 5 atm. Results show that by increasing p, not only values of SL decrease but also lean flammability limits can be expanded. It is found that the effect of p on SL is more sensitive when ? is close to the lean flammability limit than near the stoichiometry (? = 1). As to the effect of hydrogen compositions, the higher hydrogen concentration is, the higher value of SL and the wider range of lean flammability limits. On the other hand, the effect of p on SL is diminished at higher hydrogen concentrations.
For turbulent premixed syngas flames, we measure high-pressure ST of lean premixed syngas mixtures (35%H2/65%CO) at ??= 0.5 covering a wide range of pressure from p = 0.1 MPa to 1.0 MPa. Note that the present measurements differ with previous studies, because we keep the turbulent Reynolds numbers (ReT ? u?LI/?) constant by adjusting the root-mean-square turbulent fluctuation velocity (u?) and the integral length scale (LI) in proportion to the decreasing kinematic viscosity of reactants (?) at elevated pressure. Results show that, contrary to popular scenario for turbulent flames, when ReT is fixed, ST decreases similarly as SL with increasing p in minus exponential manners, ST ~ p-n, where n ≈ 0.5 almost independent of ReT. Moreover, at any fixed p, values of ST/SL increase noticeably with increasing ReT. It is found that these very scattering ST/SL data over very wide ranges of p and ReT can be merged onto a single curve having a relation of ST/u? = 0.49Da0.25, where Da is the turbulent Damkohler number. This experimental relation supports a theoretical prediction proposed by Zimont. However, such relation is different from our previous finding for lean methane mixtures where ST/u? = 0.14Da0.5 using the same facility and at the same p and ReT. We shall discuss the effects of pressure, flame instability and turbulence scale in attempt to address the aforementioned discrepancy in the exponential power constants of the Da dependence between lean syngas and methane mixtures. These results should be important when applying syngas fuels with higher hydrogen compositions to gas turbines, industrial furnaces, and other various burners.

並列關鍵字

lean high-pressure premixed turbulent combustion ； lean flammability limit ； laminar and turbulent burning velocities ； Hydrogen/carbon monoxide syngas ； turbulent Damkohler number. ； turbulent Reynolds number

參考文獻

[61] 邱建文，“氫氣/一氧化碳合成氣於高壓層流與紊流環境下之燃燒速度量測”，國立中央大學機械工程研究所，碩士論文，2011年。

[8] 李尚軍，“火花引燃機制與散佈狀燃燒形態之實驗研究”，國立中央大學機械工程研究所，碩士論文，2008年。

[7] 石偉達，“預混紊流燃燒：火花引燃機制與加氫效應之定量量測”，國立中央大學能源工程研究所，碩士論文，2007年。

[20] Coward, H. F. and Jones, G. W., “Limits of flammability of gases and vapors“Rep. Invest. U.S. Bur. Min. No.503,1952

[1] The Department of Energy (DOE). http://www.doe.gov/.