Recognition of leak CO₂ with wavelet analysis based on correlation monitoring between CO₂ and O₂ in atmosphere

It is necessary to monitor and identify CO₂ leakage from geosequestration project. Some experiments were designed to test the feasibility of characterization and recognition method for CO₂ leakage based on correlation analysis between atmospheric CO₂ and O₂. The monitoring results showed that atmospheric CO₂ had a good linear relationship with O₂, especially in the day time. The variations of CO₂ and O₂ were influenced by the weather condition. The results of emission experiments proved the capability of coupling monitoring atmospheric CO₂ and O₂ to recognize the abnormal CO₂ signal due to leakage. Moreover, two parameters, ALF (apparent leakage flux) and ALFs (apparent leakage flux based on surplus variable) were used to distinguish CO₂ leakage in real time. The monitoring results verified that it was feasible to identify CO₂ leakage with these two parameters. Compared with the results after median filter, the abrupt leakage signal can be more easily recognized with that after wavelet filter. ALFs can be applied to estimate the elevated concentration due to leakage, which can also be predicted by interpolating with related variation of atmospheric O₂. ALF reflects the relative leakage rate. Therefore, it is practicable to identify leakage event in real time with the method proposed in this paper.