-
本文中主要采用TDLAS气体测量原理和代数重建算法来实现对甲烷扩散区域的2维体积分数分布重建。
-
TDLAS气体测量应用朗伯比尔定律[16]:
$ \frac{{{I_{\rm{t}}}}}{{{I_0}}} = {\rm{exp}}( - {k_\nu }L) $
(1) 在单一跃迁前提下可以写作:
$ \frac{{{I_{\rm{t}}}}}{{{I_0}}} = {\rm{exp}}[-pxS\left( T \right){\varphi _\nu }L] $
(2) 式中, It为透射光强(mW),I0为入射光强(mW);kν为光谱吸收系数(cm-1),由气体的静态压强p(Pa)、线强S(T)(cm-2·Pa-1)、线型函数φν和体积分数x确定;ν为入射光频率(cm-1),L为有效吸收光程(cm),φν为归一化的线型函数,其积分值为1。
对(2)式进行积分,进而忽略线型函数的影响,可以得到积分吸光度A和体积分数x之间的关系,即:
$ A = \int {{\rm{ln}}} \left( {\frac{{{I_{\rm{t}}}}}{{{I_0}}}} \right)d\nu = pS\left( T \right)xL $
(3) 得到一个包含积分吸光度A、体积分数x、光程L的方程组。从而可以利用ART方法进行迭代计算。
-
ART代数重建算法是迭代重建法的一种形式,它的特点是:预设一初始图像x0,以x0为基础计算x1,进而根据x1计算x2,重复迭代至满足预设收敛条件,而后终止。每一次通过xk计算xk+1时,只需考虑一条光线产生的校正值Δxk,其所修正的数据网格也只包含这条光线通过的网格[17-20]。其本质上是通过一条条光线对所选初值的迭代修正,逐步令重建结果不断逼近实际值,在迭代修正过程中每一条光线并不总能起到正面的作用,在实验数据不够稳定的情况下,加入一些误差较大的光线将会放大误差,降低重建效果,对于重建算法中光线的选取是一个比较重要的环节,并且与实际测量的体积分数场分布及光线分布有着紧密的联系。
将(2)式改写为:
$ \frac{A}{{pS(T)}} = {R_{M \times N}}x $
(4) 式中,A为实验数据处理后获得的积分吸光度向量,x为所求体积分数向量,而矩阵R则为投影系数矩阵。
采用ARTⅡ算法,则有:
$ {x_{k + 1}} = \left\{ \begin{array}{l} {x_k},({\rm{ }}a{_{{i_k}}} \ge {\rm{ }}{r_{{i_k}}}^{\rm{T}}{x_k})\\ {x_k} + \lambda \cdot \frac{{{a_{{i_k}}} - {r_{{i_k}}}^{\rm{T}}{x_k}}}{{\parallel {r_{{i_k}}}{\parallel ^2}}} \cdot {r_{{i_k}}},\left( {other} \right) \end{array} \right. $
(5) 式中,ik=k(modI)+1,k代表迭代次数,I是R的维数,λ为松弛因子,且取值在(0,2]之间,aik为积分吸光度A的分量,rik为投影系数矩阵R的分量,经过k次迭代满足截止条件时的xk即为重建结果。算法流程图如图 1所示。
-
本实验中采用如图 5所示装置。将用于模拟甲烷泄漏的平面炉放置在旋转台的中心,使用电磁流量计控制气流速度,在旋转台上正交方向布置两组光线准直器及探测器,每组6对,通过旋转至0°, 45°, 90°, 135°分别采样,共获取24路信号,激光器发出的光经由分束器连接到各个准直头,其中一路用于同步定标。实验设计在平面炉的中心点右上放置透光的石英柱,对比是否放置石英柱的重建结果来与仿真实验相互验证。
实验中使用Nanoplus公司的分布式反馈激光器(distributed feedback laser,DFB),其中心波长覆盖到甲烷吸收峰位1653.72nm,在常温下吸收强度良好。对采集到的信号进行处理,计算出,经过查表获取常温下的线强值为1.097×10-8cm-2·Pa-1,压强p=1.01×105Pa,可以计算出的值,然后将其作为迭代信息进行迭代重建。
TDLAS信号检测采用直接测量法,获取信号如图 6所示。选取信号差异区域(长约500个点)前后各1000个点,利用干涉定标信号确定横坐标,然后计算积分吸光度A。
重建结果如表 1和图 7所示。表 1中上下两部分的6×6数据表格分别对应图 7a、图 7b各网格体积分数值。
volume fraction of methane 0 0.0002613 0.002157 0 0.003408 0 volume fraction distribution of Fig. 7a 1.199×10-6 0.01037 0 0.003102 1.638×10-6 0 0.005740 0 0.01137 0.0006311 0 0 1.382×10-6 0.0009636 0.006911 0 0.0006864 0.002332 3.653×10-6 0.004569 0 0.002429 0.0001081 0.001578 0.005721 3.653×10-6 1.382×10-6 0 0.001555 0.0004072 volume fraction of methane 0.0002642 0.0002642 0.0002642 0.007122 0.01247 0.007085 volume fraction distribution of Fig. 7b 0.01339 0 0 0.006233 0.006762 0 0.003168 0.01182 0.01704 0.01841 0 0.004573 0.01162 0.01373 0.004707 0 0 0.005895 0.005415 3.391×10-5 0 0.0003252 0.008037 0 0 3.610×10-5 0.01378 9.827×10-5 0 0 从图 7中的重建结果可以看出,平面炉出气并不均匀。图 7a、图 7b中已用方框标注出石英柱摆放区域。从图 7b、图 7d可以看出,在放置石英柱之前,甲烷主要集中在中心区域附近。从图 7a、图 7c可以看出,放置石英柱挡住了右侧之后,重建结果显示出右侧出现空穴,并对周边甲烷体积分数产生了影响,降低了右侧区域整体体积分数,左侧区域体积分数变化不大,与实验前预测相吻合。实验中遇到的主要问题在于气流的不稳定性,通过大量的重复试验来获取平均数据并不能完全克服这一问题,需采取增加单次扫描光线数量和提高气体体积分数缓和误差的方法改进实验,在考虑成本的前提下提高测量的准确度和重复性。
基于TDLAS检测技术的甲烷体积分数场重建研究
Research of methane volume fraction field reconstruction based on tunable diode laser absorption spectroscopy detection technology
-
摘要: 为了实现对甲烷体积分数场的2维分布重建,基于可调谐半导体激光吸收光谱(TDLAS)检测技术,以甲烷为目标气体,采用直接吸收的测量方式,探测了甲烷氮气混合气的吸收光谱信号,通过代数重建算法对甲烷体积分数进行了模拟重建和实验研究,模拟重建采用了6×6共36个方格的正方形重建区域,假定一个方形区域内具有空穴的体积分数分布,模拟24条光束从4个方向穿过重建区域,获取了模拟光线下的投影值。结果表明,经过重复实验统计均方根误差在2.58%,对模拟投影信号加入不同信噪比(5%,10%,20%)的高斯白噪声之后再进行重建,均方根误差分别在4.17%~9.30%之间;实验研究采取面源泄露式扩散方式,并通过在中心附近放置石英柱的方式人为制造体积分数空穴,形成非均匀体积分数场,通过对放置障碍物前后的重建结果对比,能够看到在空穴位置有明显的体积分数下降。TDLAS技术与计算机断层重建技术在气体体积分数场的局部分布检测上有可行性,具备作为有毒有害气体云团体积分数分布检测手段的潜力。
-
关键词:
- 光谱学 /
- 2维分布重建 /
- 代数重建算法 /
- 可调谐半导体激光吸收光谱检测技术
Abstract: In order to realize 2-D distribution reconstruction of methane volume fraction field, based on tunable diode laser absorption spectroscopy (TDLAS) detection technology, methane was used as the target gas and the absorption spectrum signal of methane and nitrogen mixture was detected by direct absorption method. Volume fraction of methane was simulated and reconstructed by algebraic reconstruction algorithm. During simulation reconstruction, square reconstruction area with total 36 squares of 6×6 was used. By assuming volume fraction of a hole in a square area and simulating 24 beams through the reconstruction region from 4 directions, the projection values under the simulated ray were obtained. The results show that root mean square error of the repeated experiment is 2.58%. After adding Gaussian white noise with different signal-to-noise ratio (5%, 10%, 20%) to the simulated projection signal, root mean square error is between 4.17%~9.30%. Surface source leakage diffusion was adopted in the experiments. Artificial volume fraction hole was created and inhomogeneous volume fraction field was formed after quartz a column was placed near the center. By comparing the reconstruction results before and after the placement of quartz column, it can be seen that there is significant decrease of volume fraction in hole position. It is feasible to detect local distribution of gas volume fraction by TDLAS technology and computer tomography, and has the potential to detect volume fraction of toxic and harmful gas clouds. -
volume fraction of methane 0 0.0002613 0.002157 0 0.003408 0 volume fraction distribution of Fig. 7a 1.199×10-6 0.01037 0 0.003102 1.638×10-6 0 0.005740 0 0.01137 0.0006311 0 0 1.382×10-6 0.0009636 0.006911 0 0.0006864 0.002332 3.653×10-6 0.004569 0 0.002429 0.0001081 0.001578 0.005721 3.653×10-6 1.382×10-6 0 0.001555 0.0004072 volume fraction of methane 0.0002642 0.0002642 0.0002642 0.007122 0.01247 0.007085 volume fraction distribution of Fig. 7b 0.01339 0 0 0.006233 0.006762 0 0.003168 0.01182 0.01704 0.01841 0 0.004573 0.01162 0.01373 0.004707 0 0 0.005895 0.005415 3.391×10-5 0 0.0003252 0.008037 0 0 3.610×10-5 0.01378 9.827×10-5 0 0 -
[1] LAI S L, WANG H, GONG P, et al. Research of propane detecting based on near-infrared laser spectral absorption[J]. Laser Technology, 2017, 41(2):284-288(in Chinese). [2] KRISHNA Y, O'BYRNE S. Tunable diode laser absorption spectroscopy as a flow diagnostic tool:A reviewreview[J]. Journal of the Indian Institute of Science, 2016, 96(1):17-28. [3] NADIR Z, BROWN M S, COMER M L, et al. Gaussian mixture prior models for imaging of flow cross sections from sparse hyperspectral measurements[C/OL].(2015-12-16)[2018-02-01].https://www.researchgate.net/publication/300413419_Gaussian_mixture_prior_models_for_imaging_of_flow_cross_sections_from_sparse_hyperspectral_measurements. [4] LINDSTROM C, TAM C J, DAVIS D, et al. Diode laser absorption tomography of two dimensional supersonic flow[C/OL].(2007-07-11)[2018-02-01].https://www.researchgate.net/publication/268482107_Diode_Laser_Absorption_Tomography_of_Two_Dimensional_Supersonic_Flow. [5] MASTERSON B, SAPPEY A, OWENBY D, et al. Flight-ready TDLAS combustion sensor for hypersonics[C/OL].(2007-07-11)[2018-02-01].https://www.researchgate.net/publication/271366737_Flight-Ready_TDLAS_Combustion_Sensor_for_Hypersonics. [6] BRYNER E, BUSA K, MCDANIEL J, et al. Spatially resolved temperature and water vapor concentration distributions in a flat flame burner by tunable diode laser absorption tomography[C/OL].(2011-01-07)[2018-02-01]. http://www.researchgate.net/publication/268566152_Spatially_Resolved_Temperature_and_Water_Vapor_Concentration_Distributions_in_a_Flat_Flame_Burner_by_Tunable_Diode_Laser_Absorption_Tomography. [7] DONG L, TITTEL F K, LI C, et al. Compact TDLAS based sensor design using interband cascade lasers for mid-IR trace gas sensing[J]. Optics Express, 2016, 24(6):A528-A535. doi: 10.1364/OE.24.00A528 [8] DING N, ZHANG S B. Water vapor tomography algebraic reconstruction algorithm based on the prime number decomposition access order[J]. Geography and Geo-Information Science, 2017, 112(3):42-47. [9] JIANG Zh Sh, WANG F, XU T, et al. Study of methane telemetry based on tunable semiconductor laser absorption spectroscopy[J]. Energy Engineering, 2012(3):5-9(in Chinese). [10] MA L, LI X, CAI W, et al. Selection of multiple optimal absorption transitions for nonuniform temperature sensing[J]. Applied Spectroscopy, 2010, 64(11):1274-1282. doi: 10.1366/000370210793335052 [11] BUSA K M, RICE B, MCDANIEL J C, et al. Direct measurement of combustion efficiency of a dual-mode scramjet via TDLAT and SPIV (Invited)[C/OL].(2015-01-09)[2018-02-01].http://www.researchgate.net/publication/306033342_Direct_Measurement_of_Combustion_Efficiency_of_a_Dual-Mode_Scramjet_via_TDLAT_and_SPIV_Invited. [12] LIU C, XU L, CHEN J, et al. Development of a fan-beam TDLAS-based tomographic sensor for rapid imaging of temperature and gas concentration[J]. Optics Express, 2015, 23(17):22494. doi: 10.1364/OE.23.022494 [13] LI J Y. Study on temperature measurement of combustion field based on tunable laser absorption spectroscopy[D]. Tianjin: Tianjin University, 2013: 1-135(in Chinese). [14] SONG J L, HONG Y J, WANG G Y, et al. Study on two-dimensional distribution reconstruction of gas temperature and concentration in combustion field based on laser absorption spectroscopy[J]. Acta Physica Sinica, 2012, 61(24):240702(in Chinese). [15] CHENG L H. Study of iterative tomography algorithm for gas diffusion and distribution reconstruction based on incomplete projection data[D]. Hefei: Hefei Polytechnic University, 2015: 1-78(in Chin-ese). [16] XIA H H, KAN R F, LIU J G, et al. Analysis of algebraic reconstruction technique for accurate imaging of gas temperature and concentration based on tunable diode laser absorption spectroscopy[J]. Chinese Physics, 2016, B25(6):246-253. [17] BUSA K M, BROWN M S, GRUBER M, et al. Common-path measurement of H2O, CO, and CO2 via TDLAS for combustion progress in a hydrocarbon-fueled scramjet[C/OL].(2016-01-08)[2018-02-01].http://www.researchgate.net/publication/309339754_Common-Path_Measurement_of_H2O_CO_and_CO2_via_TDLAS_for_Combustion_Progress_in_a_Hydrocarbon-Fueled_Scramjet. [18] XIA H, XU Z, KAN R, et al. Numerical study of two-dimensional water vapor concentration and temperature distribution of combustion zones using tunable diode laser absorption tomography[J]. Infrared Physics & Technology, 2015, 72:170-178. [19] ZHANG G, LIU J, XU Z, et al. Characterization of temperature non-uniformity over a premixed CH4-air flame based on line-of-sight TDLAS[J]. Applied Physics, 2016, B122(1):3-11. [20] JACKSON K, GRUBER M, BARHORST T. The HIFiRE flight 2 experiment: an overview and status update[C/OL].(2009-08-05)[2018-02-01]. http://www.researchgate.net/publication/268482795_The_HIFiRE_Flight_2_Experiment_An_Overview_and_Status_Update.