介绍了可调谐半导体激光吸收光谱技术(TDLAS)原理和实验系统, 并对系统噪声进行了分析; 以体积比浓度为90×10-6和30×10-6的NH3为例, 利用TDLAS系统采集了该浓度气体的二次谐波原始光谱。 为改善光谱信号, 分别用五种数字滤波方法对原始光谱进行了滤波处理比较, 做了NH3的浓度梯度实验并对浓度为20×10-6 NH3进行了长时间监测实验。 实验结果表明, 算术平均-小波变换滤波相比其他方法更有效地对原始光谱信号进行了改善, 提高了系统信噪比和信号平滑度, 使系统浓度检测限由原来的10×10-6降低到1.25×10-6, 信噪比提高了约14倍, 为逃逸氨极低浓度检测提供了一种较为有效的数据预处理方法。

Tunable diode laser absorption spectroscopy technology (TDLAS), with its advantages of high selectivity and accuracy, provides a reliable approach to the on-line detection of escaping ammonia. Firstly, the present paper introduces the TDLAS principle, experimental system and the analyses of system noise. Then with the concentration of 90×10-6 and 30×10-6 NH3 for example, we used TDLAS system to collect their second harmonic original spectrum with all kinds of noise interference. To improve the signal spectrum, five types of digital filtering methods were respectively used to filter the original spectrum. Finally we did the NH3 experiments of concentration gradient and the long time monitoring: NH3 experiment of 20×10-6. The analysis indicated that the averaging-wavelet filtering is validated to be more accurate than the other filtering methods in the noise reduction, which can improve the precision of the monitoring system from 10×10-6 to 1.25×10-6 and the SNR also increases by 14 times. It provides an effective pretreatment during the monitoring of escaping ammonia of extremely low concentration.