Quantitative analysis of reaction gases or exhaust using an online process mass spectrometer

Online quantitative analysis of reaction gases or exhaust in industrial production is of great significance to improve the production capacity and process. A novel method is developed for the online quantitative analysis of reaction gases or exhaust using quantitative mathematical models combined with the linear regression algorithm of machine learning. After accurately estimating the component gases and their contents in the reaction gases or exhaust, a ratio matrix is constructed to separate the relevant overlapping peaks. The ratio and calibration standard gases are detected, filtered, normalized, and linearly regressed with an online process mass spectrometer to correct the ratio matrices and obtain the relative sensitivity matrices. A quantitative mathematical model can be established to obtain the content of each component of the reaction gases or exhaust in real time. The maximum quantification error and relative standard deviation of the method are within 0.3% and 1%, after online quantification of the representative yeast fermenter tail gas.