Flow regime identification methodology with neural networks and two-phase flow models

Mi, Ye; Ishii, Mamoru; Tsoukalas, Lefteri H.

doi:10.1016/s0029-5493(00)00325-3

Cited by 151 publications

(64 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Finally, the area averaged void fraction can be obtained from the measured admittance. Cross-calibration made with gamma densitometer shows that the void fraction is almost a linear function of the normalized admittance (Mi et al, 2001b). In addition, to convert from the measured admittance to the area-averaged void fraction value a calibration is made using differential pressure measurements at low liquid velocities (Mi et al, 2001a).…”

Section: Impedance Metermentioning

confidence: 99%

“…A significant advance in the objective flow regime mapping was achieved by the use of artificial neural networks (ANN) (Cai et al, 1994;Mi et al 1998;2001a, 2001b. Using statistical parameters from the void fraction distribution and Kohonen Self-Organizing Neural Networks (SONN) it was possible to identify the flow regimes more objectively.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On the use of area-averaged void fraction and local bubble chord length entropies as two-phase flow regime indicators

et al. 2010

View full text Add to dashboard Cite

show abstract

Section: Impedance Metermentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

On the use of area-averaged void fraction and local bubble chord length entropies as two-phase flow regime indicators

et al. 2010

View full text Add to dashboard Cite

show abstract

“…On the other hand, the patterns of cross-sectional average electric capacitance of the flow field 15,16 and the variations and fluctuations of local absolute and differential pressures [17][18][19][20][21][22][23] , for example, have been found to be strongly regime-dependent. While the temperature effect when using the impedance technique still needs to be overcome 24 , the pressure fluctuations that result from the passage of gas and liquid pockets, and their statistical characteristics, are particularly attractive because the required sensors are robust, inexpensive, and relatively well-developed, and are thus more likely to be applied in the industrial systems 25 .…”

Section: Introductionmentioning

confidence: 99%

“…To avoid subjective judgment, artificial neural network (ANN) modeling and statistical analysis methods have been employed to implement non-linear mappings from measurable physical parameters to flow regimes 16,24,26 . Artificial neural networks are analytical tools that imitate the neural aspect of the human brain, whereby learning is based on experience and repetition rather than the application of rule-based principles and formulas.…”

Section: Introductionmentioning

confidence: 99%

“…Mi et al 16,24 applied a neural network for two-phase flow regime identification in a vertical channel using signals from electric capacitance probes with excellent results. Gupta et al 27 successfully applied a hybrid method based on four neural networks along with simple first-principle models (the latter intended to render the model independent of specific device geometry), for prediction of the attachment rate constant in flotation columns.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Bubble Size Control to Improve Oxygen-Based Bleaching: Characterization of Flow Regimes in Pulp-Water-Gas Three-Phase Flows

Karrila¹

2006

View full text Add to dashboard Cite

SUMMARYThis document reports on the results of a research program aimed at the elucidation of various hydrodynamic aspects of low consistency paper pulp-water-gas three-phase flows, conducted jointly by the G.W. Woodruff School of Mechanical Engineering, Georgia Tech, and the Institute of Paper Science and Technology.Hydrodynamic flow characteristics of solid-liquid-gas slurry made by intimately mixing fibrous paper pulp with water and air were first investigated in a short, vertical circular column.The pulp consistency (weight fraction of pulp in the pulp-water mixture) was varied in the low consistency range of 0.0 ~ 1.5%. The test section was 1.8 m long, with 5.08 cm inner diameter.Mixing of the slurry prior to entering the test section was done using a patented mixer with controlled cavitation that generated finely dispersed micro-bubbles. Flow structures, gas holdup (void fraction), and the geometric and population characteristics of gas bubbles in the gas-pulpliquid three-phase flow were experimentally investigated, using visual observation, Gamma-ray densitometry, and flash X-ray photography. Superficial velocities of the gas and liquid/pulp mixture covered the ranges 0~26 cm/s and 21~51 cm/s, respectively.Five distinct flow regimes could be visually identified. These included dispersed bubbly, characterized by isolated micro-bubbles entrapped in fiber networks; layered bubbly, characterized by bubbles rising in a low consistency annular zone near the channel wall; plug; churn-turbulent; and slug. The dispersed and layered bubbly regimes could be maintained only at very low gas superficial velocities or gas holdups. Flow regime maps were constructed using phasic superficial velocities as coordinates, and the regime transition lines were found to be sensitive to consistency.The cross section-average gas holdup data showed that both the dispersed and the layered bubbly regimes could best be represented by the homogeneous mixture model. The drift flux model could best be applied to the reminder of the data when the plug and churn-turbulent flow regimes were treated together, and the slug flow was treated separately. The drift flux parameters depended on the pulp consistency.The feasibility of using artificial neural networks (ANNs) for the identification of the flow regimes in the aforementioned three phase gas/liquid/pulp fiber systems, using the statistical characteristics of pressure fluctuations measured by a single pressure sensor, was then examined.The aforementioned experimental test facility was used for this purpose. Experiments were carried out in the transparent, vertical circular column. The pulp consistency (weight fraction of 2 pulp in the pulp-water mixture) was varied in the low consistency range of 0.0 ~ 1.5%, and flow structures and other hydrodynamic characteristics were experimentally investigated. Local pressure fluctuations at a station 1.2 m above the test section inlet were also recorded using a highsensitivity pressure transducer that was installed in a manner not to cause flow distu...

show abstract

ANN–based prediction of two‐phase gas– liquid flow patterns in a circular conduit

2006

View full text Add to dashboard Cite

Flow regime identification methodology with neural networks and two-phase flow models

Cited by 151 publications

References 16 publications

On the use of area-averaged void fraction and local bubble chord length entropies as two-phase flow regime indicators

On the use of area-averaged void fraction and local bubble chord length entropies as two-phase flow regime indicators

Bubble Size Control to Improve Oxygen-Based Bleaching: Characterization of Flow Regimes in Pulp-Water-Gas Three-Phase Flows

ANN–based prediction of two‐phase gas– liquid flow patterns in a circular conduit

Contact Info

Product

Resources

About