气体成分与弛豫信息对应关系的声谱盲分解研究

Gas acoustic absorption spectrum can be obtained from the acoustic energy absorption versus frequencies of acoustic wave, while the gas optical absorption spectrum results from the electromagnetic wave radiation. Whether gas acoustic spectrum can, like optical spectrum, provide unique feature to recognize gas species is an interdisciplinary problem on acoustics and signal processing with great application prospect. Gas acoustic absorption spectrum essentially results form gas acoustic relaxation, which is caused by gas molecular vibration and energy level transition. Since each kind of molecule gas has unique vibrational mode and energy level transition mode, the acoustic spectrum of each kind of molecule gas will be various. So it could be concluded that the acoustic spectrum of a gas mixture will contain all the acoustic relaxation information of its compositions. For all gas mixtures with polyatomic molecules, the acoustic relaxing energy is released via every vibrational-translational deexcitation path, which is a one-to-one correspondence to the vibrational mode of each kind gas. Consequently, a multi-relaxation spectrum of gas can be seen as the sum of decomposed single-relaxation spectrums determining by vibrational mode. Based on these considerations, we propose a theory for detecting target gases from gas mixtures with unknown compositions by using blind decomposition and feature extraction method for gas acoustic spectrum in this project. There are two following objectives in this project: 1) aims to establish a blind decomposition theory to decompose a gas acoustic spectrum into several single-relaxation spectrums without any information of its compositions, 2) aims to establish a signal processing techniques based method to extract the features of gas acoustic spectrum, which contains the characteristics of decomposed single-relaxation spectrums. The achievements of this project would demonstrate potential for bridging the theoretical gap between acoustic absorption spectrum and optical absorption spectrum of gas.

源于机械波的声吸收谱与源于电磁波的光吸收谱是否相似，能否也从声谱上获取与气体种类一一对应的特征信息进行气体探测？是声学和信号处理交叉领域中一个极具应用前景的科学问题。声弛豫是形成气体声吸收谱的核心原因，引起声弛豫的分子振动模式及能级跃迁具有唯一性，因此每种气体的声谱具有唯一性，混合气体的声谱含有所有成分气体的声弛豫信息。本项目通过分析气体声弛豫需要分解到每一条振动-平动去激发路径释放能量的现象，研究复合弛豫声谱分解为单弛豫声谱的模型，进而重点研究将成分未知气体的声谱盲分解为单弛豫声谱的方法，同时研究从气体声谱上盲提取单弛豫声谱信息的信号处理方法，并利用振动-平动去激发路径与分子内振动模式具有的一一对应关系，寻找盲分解和提取的单弛豫声谱与气体种类一一对应的关系，最终找到对应气体种类的唯一"声谱信息识别码"，实现从未知背景中探测目标气体，完成声气体探测的理论突破，弥补气体声谱与光谱的理论差距。

声弛豫和光弛豫分别是形成气体声吸收谱和光吸收谱的核心原因。引起弛豫的分子振动模式具有唯一性，因此声谱和光谱应该具有相同的气体探测基础。对于光谱，分子振动模式唯一吸收某一特定波长的光波。然而，声波通过波动所有分子外部自由度激励所有分子内部振动模式。所以混合气体的声谱含有所有成分气体的声弛豫信息，需要从声谱上盲分解并提取出与气体种类一一对应的特征信息完成气体探测。本项目完成了如下研究内容：.1）研究气体声弛豫的有效热容具有分解为单一弛豫过程有效热容叠加的形式，利用信号分解的思想将有效热容的分解拓展到气体弛豫声吸收谱上，理论上完整证明了气体的弛豫声吸收谱是由单弛豫过程声吸收谱在全频段叠加而成，完成了气体声谱盲分解；.2）将传统的平行弛豫理论和串行弛豫理论相结合，提出了一种新的声弛豫吸收理论模型，实现了气体成分弛豫贡献的明确解析表达，即每种气体成分在混合气体的弛豫过程中贡献一个单弛豫过程；.3）最终发现了气体内部热容随温度的相对变化率就是气体种类的弛豫“识别码”，即纯气体弛豫过程与其在混合气体中贡献的单一弛豫过程具有相同的内部热容温度相对变化率。因此混合气体单一弛豫过程的内部热容温度相对变化率能够用于气体成分的定性识别，且该内部热容与对应纯气体内部热容的比值就是气体成分的浓度体积比。从而提出了基于内部热容温度相对变化率的气体传感方法。.综上所述，能够利用本项目提出的声谱盲分解方法，从实验声谱点中快速重建出若干个单一弛豫声谱；然后根据本项目提出的声弛豫吸收理论模型，认定这些声谱属于一一对应的气体成分；最后利用求得的单弛豫过程弛豫强度（正比于气体内部热容）的温度相对变化率，探测这些气体成分的种类和浓度。从而完成了本项目的核心研究目标，并最终肯定地回答了本项目申请的科学意义：源于机械波的声吸收谱与源于电磁波的光吸收谱具有相似性；能够从声谱上获取与气体种类一一对应的特征信息进行气体探测！