Flow pattern recognition of condensing steam jet in water flow by chaotic characteristic analysis

Online flow pattern recognition of direct contact condensation (DCC) is critical for the design and stable, efficient operation of phase-change systems that utilize steam–water jet injectors or direct contact heat exchangers. In this study, based on the chaotic analysis method, the high-accuracy flow pattern recognition is realized by selecting only two feature parameters, i.e., the correlation dimension and the Hurst exponent. This method is validated across a broad range of inlet conditions: vapor mass flux (200–600 kg m−2 s−1), Reynolds number of subcooled water (7.8 × 104–17.2 × 104), and inlet subcooled water temperature (20–60 °C). The dynamic characteristics of three typical flow patterns—stable jet, oscillation jet, and bubble flow—are systematically examined. The results show that the correlation dimension values for these three flow regimes range within 11.06–13.20, 10.02–11.81, and 8.87–10.45, respectively, while the corresponding Hurst exponent values fall in the intervals of 0.293–0.483, 0.204–0.319, and 0.146–0.25. An overall accuracy of 82.40% was achieved by the identification method based on these two parameters, corresponding to individual accuracies of 83.00% for bubble flow, 80.40% for oscillation jet, and 84.70% for stable jet. The proposed approach illustrates high effectiveness and practicality, offering a reliable new method for the flow pattern recognition in DCC processes within confined channels.