Flow Regime Identification in a Bubble Column Based on Both Statistical and Chaotic Parameters Applied to Computed Tomography Data

The Kolmogorov entropy (KE) algorithm was applied successfully to single source γ‐ray Computed Tomography (CT) data measured in a 0.162 m ID bubble column equipped with a perforated plate distributor (163 holes · ∅︁ 1.32 mm). Dried air was used as the gas phase and Therminol LT (ρL = 886 kg m–3, μL = 0.88 · 10–3 Pa s, σ = 17 · 10–3 N m–1) was used as a liquid phase. Three different pressures, P, of 0.1, 0.4, and 1.0 MPa were examined. At each pressure the superficial gas velocity, uG, was increased stepwise by steps of 0.01 m s–1 up to 0.2 m s–1. The average absolute deviation (AAD) was also used as a robust statistical criterion for regime transition. At all three pressures, based on the sudden changes in both the AAD and KE values, the boundaries of the following five regimes were identified: dispersed bubble regime, first and second transition regimes, coalesced bubble regime consisting of four regions (called 4‐region flow), and coalesced bubble regime consisting of three regions (called 3‐region flow). The existence of these regimes has already been documented. As the pressure increases, the transition velocity between the dispersed bubble and first transition regimes and the transition velocity between coalesced bubble (4‐region flow) and coalesced bubble (3‐region flow) regimes shift to higher uG values. On the other hand, at P = 0.4 MPa the second transition regime starts earlier. In addition, at P = 1 MPa the transition to coalesced bubble (4‐region flow) is delayed. By means of statistical (average absolute deviation) and chaotic (Kolmogorov entropy) parameters applied for the first time to nonintrusive Computed Tomography (CT) data, the boundaries of five different regimes were identified.