基于卡尔曼滤波的激光光谱CO_2监测算法研究

Inversion algorithm of tunable diode laser absorption spectroscopy to determine CO_2 concentration based on kalman filter algorithm

目的:为了进一步抑制光谱干扰对可调谐二极管激光吸收光谱技术测量结果的影响。方法:基于搭建的可调谐激光光谱技术气体监测系统,对CO2标准气体进行了实验分析。在最小二乘法的基础上,提出了采用卡尔曼滤波算法进行建模并预测的方法,并以此预测了CO2气体的浓度。结果:与采用卡尔曼滤波算法前相比,预测值和参考值的相关系数由0.9961增大到0.9999,预测均方根误差由0.1%减小至0.0034%。结论:该方法用于TDLAS技术中,可以实现气体的准确测量。

Objective：To reduced the spectrum interference in tunable diode laser absorption spectroscopy （TDLAS） measurements. Methods： The modeling experiment of CO2 standard gas was implemented by the gas monitoring system. The kalman filtering algorithm modeling method was proposed based on the least squares method. CO2 gas concentrations were predicted. Results： The correlation coefficient of the prediction and the reference value is 0. 9999 compared with 0. 9961 of the non kalman filter method. The root mean square error of perdition is 0.0034% compared with 0.1% of the non kalman filter method. Conclusion： The proposed method can be adopted to invert gas concentrations in the TDLAS.