In the field of trace gas measurement, with the characteristics of high sensitivity, high selectivity and rapid detection, tunable diode laser absorption spectroscopy (TDLAS) is widely used in industrial process and trace gas pollution monitoring. Herriott cell is a common form of multiple reflections of the sample cell, the structure of the Herriott cell is relatively simple, which be used to application of trace gas absorption spectroscopy. In the pragmatic situation, the gas components are complicated, and the continuous testing process for a long time can lead to different degree of pollution and corrosion for the reflector in the sample cell. If the mirror is not cleaned up in time, it will have a great influence on the detection accuracy. In order to solve this problem in the process of harsh environment detection, this paper presents a design of the built-in sample cell to avoid the contact of gas and the mirror, thereby effectively reducing corrosion pollution. If there is optical pollution, direct replacement of the built-in optical sample cell can easily to be disassembled, and cleaned. The advantage of this design is long optical path, high precision, cost savings and so on.
Tunable diode laser absorption spectroscopy (TDLAS) is a high-resolution infrared laser absorption spectroscopy technique with a non-contact measurement, high spatial and temporal resolution, extensive measurement information, which has been a hot research area at present. Compared to traditional techniques, TDLAS technology has many advantages, but in engineering applications under complex environmental conditions, TDLAS technology is still facing many difficulties. Because of the impact of environmental factors, the measured spectral signal would be distorted, and cannot be used to extract useful information. Therefore, to extract useful information from the raw signal, it is essential to improve the signal to noise ratio. To eliminate interference information contained in the spectral signal, the absorption spectra of the laboratory intends to take data preprocessing methods. In the preprocess, the Empirical Mode Desperation (EMD) method is developed in recent years, which is a new self-adaptive local frequency analysis method. Compared to the method of wavelet denoising, EMD method with adaptive filters is able to achieve a multi-scale decomposition of the noise signal. In this paper, EMD method is taken to eliminate noise and interference signal source decomposition. By reconstructing the actual signal and eliminating the noise components, a better SNR can be achieved.