Multi-Block Integrated Monitoring of Gas-Liquid Two-Phase Flow Based on Conductance Signal Analysis

The state monitoring of gas-liquid two-phase flow with complex and diverse structures is the key to ensure the safe and stable operation of related industrial production. In order to achieve effective monitoring with limited sensing equipment, a multi-block integrated monitoring strategy for gas-liquid two-phase flow based on conductance signal analysis in time, frequency and multiscale domain is proposed. Three-block information including water holdup in time-domain, continuous wavelet transforms (CWT) based coarse scale approximation (CSA) coefficients in frequency-domain and CWT based local wavelet energy (LWE) coefficient maps in multiscale-domain are used for block-based modeling and monitoring. Considering the original structure of each block, the traditional principal component analysis (PCA) is adopted to model and monitor 1-D information including water holdup and CSA coefficients, while the tensor PCA is adopted for the 2-D LWE coefficient map. Three-block monitoring results are integrated through Bayesian inference to realize final monitoring. The dynamic experiment results demonstrate that the proposed multi-block integrated monitoring outperforms single-block information monitoring.