Wang Hui-feng, Jiang Xu-qian
School of Electronic & Control Engineering, Chang'an University, Xi'an 710064, China.
Guang Pu Xue Yu Guang Pu Fen Xi. 2012 Jan;32(1):171-4.
The key and challenge problem of in-situ monitoring poisonous elements of gases is how to separate the various gases absorption signal from mixed gases absorption spectroscopy and compute it's accuracy concentration? Here we present a new algorithms in return recursion iteration based on Lambert-Beer principle. In the algorithms, recurred by the character of absorption peak of various gases in the band of 190-290 nm UV rays continuous spectroscopy and the character of twin element fold for absorbance are used. Firstly, the authors suppose that there is no absorption for others gases in the character absorption band for a certain gas, the authors can inference the initial concentration of the gas. Then the authors switch to another character spectroscopy, and put the photons that gases absorption out of the total number of absorbed photons that are measured. So we could get the initial concentration of another gas. By analogy the authros can get the initial concentration of all kinds of other poisonous elements. Then come back to the character spectroscopy of the first gas, the authors can get a new concentration of the first gas from the difference between the total number of absorbed photons and the photons that other gases absorption. By analogy the authors can get the iterative concentration of other gases, by irterating this process repeatly for some times until the measurement error of the adjacent gas concentration is smaller than a certain numerical value. Finally the authors can get the real and accurate concentration of all kinds of gases. Experiment shows that the authors can get the accurate concentration of all kinds of gases with the algorithm. The accuracy can be within 2%, and at the same time, it is easy enough to satisfy the necessity of real-time requirement. In addition it could be used to measure the concentration of many kinds of gas at a time. It is robust and suitable to be taken into practice.
气体中有毒元素原位监测的关键和挑战性问题在于如何从混合气体吸收光谱中分离出各种气体的吸收信号并计算其准确浓度？在此，我们提出一种基于朗伯-比尔原理的反向递归迭代新算法。在该算法中，利用了190 - 290 nm紫外线连续光谱波段中各种气体吸收峰的特征以及吸光度的双元素折叠特征。首先，假设在某一气体的特征吸收波段内其他气体无吸收，从而推断该气体的初始浓度。然后切换到另一种特征光谱，将气体吸收的光子从测量的总吸收光子数中去除，进而得到另一种气体的初始浓度。以此类推，可得到各种其他有毒元素的初始浓度。接着回到第一种气体的特征光谱，根据总吸收光子数与其他气体吸收的光子数之差得到第一种气体的新浓度。以此类推，可得到其他气体的迭代浓度，重复此过程若干次，直到相邻气体浓度的测量误差小于某一数值。最终可得到各种气体的真实准确浓度。实验表明，利用该算法能够得到各种气体的准确浓度，精度可达2%以内，同时易于满足实时性要求，此外还可一次性测量多种气体的浓度，该算法稳健且适合实际应用。
