Underwater bubble escape volume measurement based on passive acoustic under noise factors: Simulation and experimental research

Passive acoustic methods are suitable for long-term seabed gas leak monitoring with high accuracy under the condition of low cost and simple experimental equipment. However, the impact of noise in complex underwater environments poses great challenges for measurements. In this paper, aiming at the quantitative detection problem, a theoretical model of the bubble generation process by underwater gas escape is established, and the mechanism of bubble acoustic signal generation is clarified. The model identification of bubble volume oscil-lation is realized through the complementary ensemble empirical mode decomposition (CEEMD) method. Then continuous bubbles are segmented by the normalized energy method to achieve a high-precision measurement of volume inversion. A measurement prototype is developed and tested in the lake with multiple noises caused by seawater disturbance, sea wind, and ships. The results showed that the measurement errors could be kept within 10% even under the condition of strong noise interference.

Automated summary

Passive acoustic methods are suitable for long-term seabed gas leak monitoring with high accuracy under the condition of low cost and simple experimental equipment. However, the impact of noise in complex underwater environments poses great challenges for measurements. In this paper, a theoretical model of the bubble generation process by underwater gas escape is established, and the mechanism of bubble acoustic signal generation is clarified. The model identification of bubble volume oscillation is realized through the complementary ensemble empirical mode decomposition (CEEMD) method. Then continuous bubbles are segmented by the normalized energy method to achieve a high-precision measurement of volume inversion. A measurement prototype is developed and tested in the lake with multiple noises caused by seawater disturbance, sea wind, and ships. The results showed that the measurement errors could be kept within 10% even under the condition of strong noise interference.